![Measuring SPECjbb2015 Virtualization Overhead &
Application Workloads in a Microsoft Hyper-V Cloud
John M. Medellin*, Frank C. Lee, Lokesh Budhi, Spandana Gennepally
Faculty and Graduate Assistants at the Master of Science in Information Systems Program
University of Mary-Hardin Baylor
Belton, TX 76513-2599, USA
Abstract— Virtualized environments are a valuable
component of most modern automation strategies. They
allow for creation of logical systems that can be extended
to accommodate a variety of needs in the computing
community. The critical manager of this flexibility is the
hypervisor software; it controls the direction and
limitations of the resources available to execute workloads.
Hypervisors are available in large varieties both from
reputable vendors and the open source community. This
paper executes experiments within the Microsoft Stack
(Windows 10, Windows Server 2012 R2 and Windows
Hyper-V 2016) and measures the incremental overhead
added by the virtualization layers.
The flexibility afforded by Hypervisors is not
delivered without increase in computing cost. In addition
to operating the internals of the Hypervisor software, there
are also penalties to pay for virtualization of resources
themselves. Virtualization overhead has been studied in a
variety of formats but lately, it has focused on determining
overhead driven by the impact of workloads on key
components; data, network, memory or computation, and
how different Hypervisors perform under certain types of
applications.
Our paper measures two types of workloads that
are common in the retail industry; online shopping
(business to consumer or “B2C”) vs shopping at the store.
We use the SPECjbb2015 benchmark suite to configure
different workload characteristics under each scenario
and determine workload and overhead effects using the
standard Microsoft stack. Our conclusions indicate that
there are differences in overhead incurred by the
virtualization layers and between the application
workloads themselves. There seems to be an incremental
combined overhead/workload penalty for more complex
applications. The preliminary conclusion is that this
virtualized environment gives better performance
benchmarks to online shopping than to traditional store
level sales based on the different complexities in the
applications.
Keywords— Hypervisors, B2B/B2C, Application
Workloads, Retail Applications, SPEC Corporation,
Microsoft Windows 10, Microsoft Windows Server 2012
R2, Microsoft Hyper-V 2016, Services Oriented
Architecture, SOA
I. INTRODUCTION
Virtualized environments were first introduced in the
1970s as a means to extend processing cycle usage and
avoid unacceptable levels of idle CPU resources. This
concept has been further exploited in the past decade with
the advent of Cloud environments. In the Cloud, we are able
to model a variety of logical configurations that can help us
achieve certain workloads depending on their characteristics
[11]. We can adapt our architectures to take advantage of
certain configurations that can lend better performance to
different types of application complexity.
A significant amount of work has been done in measuring
the impact of workloads on the managers of Cloud
environments. These managers (Virtual Machine Managers)
are also called “Hypervisors”. Hypervisors are available
both from traditional vendors (Microsoft in the Hyper-V
product for example [5]) or on open source (like the
OpenStack product [10]). These Hypervisors can lie on the
bare-metal (right on the actual hardware) or as extensions to
Network Operating systems as in Microsoft Hyper-V being
an extension of Microsoft Windows Server 2012 Release 2.
These configurations allow the deployment of “client”
operating systems which are resident inside the virtual
machines (for example a client-based Windows 10
implementation) and provide the execution environment for
the applications themselves.
Most of the recent body of knowledge has focused on the
measurement of the various parts of the virtualization
overhead using standard workload software (these are
available suites that focus on producing a type of
environment that is stable and has been used in research
before). One of these suites is the SPECjbb2015 by
www.spec.org. This standard workload generator can
simulate a variety of transaction mixes and can enable very
complex scenarios of business environments. The system
simulates the transaction load of a supermarket chain and
can be scaled to multiple supermarkets and multiple “central
management” offices. The objective of this system is to
measure throughput under an incremental volume scaling
algorithm until the system is saturated; the system can no
longer keep up with the load provided.
Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 3
ISBN: 1-60132-458-8, CSREA Press ©](https://image.slidesharecdn.com/114227-250418134504-5ac877e8/85/Grid-Cloud-and-Cluster-Computing-and-Applications-1st-Edition-Hamid-R-Arabnia-15-320.jpg)
![The fundamental objectives of this research are to
measure the incremental overhead and workload added by
virtualization on two common retail industry application
patterns: online shopping and store shopping. We use the
SPECjbb2015 to vary the transaction mix between online
purchases and in-store purchases for each trial. We start by
installing this environment on bare metal versions of
Windows 10 and then follow by creating virtual machine
images under Windows Server 2012 R2 with the Hyper-V
2016 hypervisor. The reduced throughput reported under
each of these Cloud environments is due to the incremental
overhead from virtualization. The two differences in
workload execution architecture patterns are analyzed
through use cases, software architecture patterns and UML
activity diagrams derived from the SPECjbb2015 code.
This document is structured as follows:
• We present a summary of the related work that has
been done in this area and how we have leveraged
those conclusions on our research.
• We structure an experiment which measures the
impact of virtualization by first executing the
SPECjbb2015 on bare metal under the Windows 10
operating system and then proceed to implement that
environment under Windows Server 2012 R2
Network OS with the Hyper-V 2016 Hypervisor. The
incremental layer of abstraction is measured and
reported by the throughput achieved in the
SPECjbb2015 suite.
• We alter the transaction load of SPECjbb2015 to
create one scenario with Web Online Shopping
(“B2C”) transactions and one other with Store
(“brick and mortar”; “B2M”) sales transactions. The
differences in application throughput performance
are also reported for each scenario.
This initial research is the first step in understanding the
effects of various types of application design patterns for
Cloud applications. We believe the differences can be
explained (and in further research optimized) by better
utilization of standard design patterns in Cloud-based
applications.
II. RELATED WORK
Extensive work has been done on the effects of
virtualized environments and some of the concepts related
to creating virtual systems date back to the 1980s [13]. The
full virtual cloud concept was only possible after the
creation of software-defined systems that managed servers,
storage, network, CPU and other resources known as Virtual
Machine Monitors or more commonly as hypervisors [3].
The hypervisor establishes policies for sourcing of resources
to the various processes that are resident on a particular
infrastructure. Hypervisors are usually complemented by
other software to create a Cloud Management Platforms
(CMP) that enhance management of the virtualized
environment [6].
A. Cloud Framework and Hypervisors
Freet, Agrawal, Walker and Badr [6] provide a review of
the general characteristics of the CMP that make
hypervisors effective. They reviewed Eucalyptus,
OpenStack, CloudStack, OpenNebula, Nimbus and
Proxmox (all open source CMPs). Their analysis concludes
that OpenStack and CloudStack have over 30 times more
messages that some of their other competitors (meaning
they are more top of mind in the development community).
They proceed to review the frameworks in relation to Xen,
KVM, Virtual Box and ESX/VMware and requirements for
data center virtualization and infrastructure provision. In
that study, various types of workloads are varied through
each candidate hypervisor and the throughput is reported.
We have adopted a similar workload variation approach and
report throughput statistics in our methodology.
B. Virtualization Overhead
As mentioned above, virtualization of resources can be a
costly proposition. Oi and Nakajima [9] explored the effects
of performance loads on the Xen environment vs others.
They determined that the performance of Xen could be
enhanced in a virtualized environment by adjusting cache
sizes in some applications. Although Xen was never able to
provide the throughput that Linux could, (both operating on
the hardware without other Operating Systems present) it
was shown how certain configuration changes could be
effected to drive higher throughput in the Xen hypervisor
environment. The point of varying configurations impact on
workload efficiency is conclusively illustrated in that work.
The benchmarking system used was SPECjbb2001 and the
effects of Network Interface Cards (NIC) were isolated so
the workload could be measured in memory usage and
throughput. We have also adopted the SPEC performance
suite but with an application development focus on the
results.
Al Jabry, Liu, Zhu and Panneerselvam [1] studied the
effects of hypervisor overhead along the following resource
usage: disk i/o, CPU, memory and VMM (hypervisor).
Their tests were conducted using VMware, Virtual Box
(Oracle Corporation) and Windows Virtual PC (Microsoft).
Their work benchmarked a standard load in each hypervisor
environment and used IOzone to quantify load on disk i/o,
RAMSpeed to quantify the impact on memory and
UnixBench, to indicate the effect on CPU. The residual
resource usage is attributed to the VMM (hypervisor)
resource. In their conclusion, they note that hypervisor
overhead is considerably low on the memory and disk i/o
segments while it is much higher on the CPU usage. This
4 Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 |
ISBN: 1-60132-458-8, CSREA Press ©](https://image.slidesharecdn.com/114227-250418134504-5ac877e8/85/Grid-Cloud-and-Cluster-Computing-and-Applications-1st-Edition-Hamid-R-Arabnia-16-320.jpg)
![study points to the need to choose hypervisors in relation to
the application workload needs and points to the Microsoft
solution as being the more balanced of the hypervisors
reviewed with those workloads. We incorporated the usage
of the Microsoft stack into our experiments because of the
tight integration within itself.
Chen, Patel, Shen and Zhou [2] studied virtualization
overhead across multiple VMs running under Xen in cloud
environments. They also found that the larger resource
usage was attributable to the CPU. They also propose a
series of equations that are remarkably accurate in
predicting the lateral scaling of workloads on all
components based on the observed results of the application
under study. Their recommendation for future work is to
vary the application characteristics (workload profiles) and
further fine-tune their predictive analysis. We incorporate
those recommendations by varying our workloads based on
B2C and store transaction loads.
C. Application Workload Research
Based on the research referenced, there is a significant
impact on utilization of CPU from the overhead generated
by the hypervisor. Further the impact is based on the type of
application that is operating in the virtualized environment.
NasiriGerdeh, Hosseini, RahimiZadeh and AnaLoui [8]
measured throughput degradation on Web applications
using the Faban suite (a web-based workload generator).
They simulated the behavior of heavy transactional Web
applications that tend to be very network intensive. Their
work also measured the effect on memory, disk i/o and
CPU. They concluded that a disproportionate difference
exists in CPU resources due to the translation of domain
addresses. This work further confirms that the principal
resource difference is the CPU utilization even when
workloads may be more i/o bound (the penalties associated
where in finding addresses; a CPU task, not access to the
actual addresses in the Web environment; an i/o task). We
incorporate this research by focusing on actual compute
power utilization rather than network or disk access. The
SPECjbb2015 suite is focused on exhausting the compute
resources rather than the disk (i/o) or network resources.
San Wariya, Nair and Shiwani [12] focused their research
on benchmarking three hypervisors; Windows Hyper-V,
VMWare/ESXi and Citrix Xen in three cloud games; 3D
Mark 11, Unigine Heaven and Halo. The objectives of their
study are to identify which hypervisor was better from a
cloud gaming workload perspective. The three performed
differently in each category but were mostly lead by the
VMWare product. For our purposes however, the HALO
benchmark (number of frames per second) is probably the
most predictive of workloads that are CPU bound. In this
category, Hyper-V performed 7% ahead of VMWare and
57% ahead of Citrix Xen. This was another reason for
selection of Hyper-V as the hypervisor for our test suite.
D. SOA Application Design Patterns
Standardized software architecture patterns provide a
series of guidelines for developing functionality that is
common across a variety of applications [14]. The work in
this domain has been futher elaborated in Services Oriented
Architecture (SOA) patterns that can be used to overcome
certain design dilemmas. Thomas Erl (with foreword by
Grady Booch) [4] presents a comprehensive set of
architecture patterns that are common in business. These
patterns were matched to the two use cases (B2C & store
below) and confirmed in the source code of SPECjbb2015.
Workloads can be modeled in a variety of ways in the
application architecture and each can achieve their goals in a
different way. If we however, standardize those patters with
relation to use cases, there is a greater possiblity for the
results to be extrapolated by other application researchers in
this field. This approach gives a better chance of replication
of the results for other studies relying on our work.
A frequently used SOA application design pattern for
abstracting logic from traditional application design
nomenclature is shown in Figure 1 below [4]. Other SOA
design patterns were confirmed in the implementation of the
use cases below in the application code itself.
Figure 1: from traditional design pattern to services
architecture design pattern (Erl et. al.)
E. The SPEC Benchmarking Suite
The SPECjbb2015 (Figure 4) constitutes a Business to
Consumer and Store to Consumer simulation for a
Supermarket Chain. The model can be extended to include
several supermarkets and several central offices in a variety
of virtual machine settings. The benchmarking suite can be
configured in a variety of business transaction settings so
that different business patterns can be simulated (e.g., web
sales versus physical store sales). The system is owned and
licensed by spec.org [15] which is a consortium of major IT
companies that have agreed on a set of principles to guide
the performance benchmarking process.
The system progressively injects transaction loads into
the environment until saturation is reached. A sample output
of these results is seen in Figure 3. In that graphic the
system begins to stress at around the 5,200 java Operations
Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 5
ISBN: 1-60132-458-8, CSREA Press ©](https://image.slidesharecdn.com/114227-250418134504-5ac877e8/85/Grid-Cloud-and-Cluster-Computing-and-Applications-1st-Edition-Hamid-R-Arabnia-17-320.jpg)
![Per Second (jOPS) with a range of 5K (median tolerance) to
50K (max tolerance). The system reaches saturation (min
tolerance) at around 6,700 jOPS and 60K. We report our
results using the total transactions up to saturation. Figure 4
is a graphic representation of the architecture of the system.
Figure 3: Sample SPECjbb2015 Benchmark Output
www.spec.org
Figure 4: SPECjbb2015 Architecture
www.spec.org
The SPECjbb products have been in existence since the
late 1990s and are useful because of their industry
acceptance. For example, Karlsson, Moore, Hagersten and
Wood [7] used an earlier version (SPECjbb2001) and
another application benchmark (ECPerf) to differentiate
effects of cache misses between different types of
applications.
III. EXPERIMENT DESIGN
As discussed above, experiments were designed where
the same application (SPECjbb2015) was installed on:
a) Bare metal with Windows 10
b) VMachine: Windows Server/Hyper-V/Windows 10
Two transaction variations were used; one increased the
B2C transaction types to be 90% of the sales while the
second one increased the store sales to 90% of the sales.
B. Application Architecture Patterns
The application patterns were analyzed by matching use
cases, identifying SOA design patterns and preparing
activity diagrams for each version of the application
configured. The details are reported below.
i. Use Case Analysis
Use cases are a functional decomposition tool that
illustrate the process interactions between actors in
applications [14]. The processes that we have selected in the
SPECjbb2015 suite are fairly standard and follow similar
patterns. The use case diagram for the store architecture is
similar to this one (www.UML-diagrams.org), the
“adornments” in the graphic describe the usage of artifacts.
Figure 5: Store Use Case Diagram
The use case diagram for B2C sales architecture pattern
is similar to the one below (www.UML-diagrams.org) :
Figure 6: B2C Use Case Diagram
At a high level, both use cases are similar in providing
functionality for the user to execute a purchase. However,
when inspected the code we found that different design
patterns and workloads were implemented. Most of these
differences are due to the inventory on-hand function at the
physical store. If the system detects an out-of-stock
condition, it will proceed to cancel and back out the
transaction. This process will take more resources than
fulfilling the order from other stores (as it does in the B2C
scenario) because it has to place the order items into
inventory and invalidate the order itself (error exception in
the store UML activity diagram below).
ii. SOA Design Patterns in SPECjbb2015
We inspected the code again to determine which patterns
were evident. Even though the processes are very similar
6 Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 |
ISBN: 1-60132-458-8, CSREA Press ©](https://image.slidesharecdn.com/114227-250418134504-5ac877e8/85/Grid-Cloud-and-Cluster-Computing-and-Applications-1st-Edition-Hamid-R-Arabnia-18-320.jpg)
![from a functional perspective, they are less similar from a
pattern design perspective. The SOA design patterns that
were used (in our opinion) are as follows (from Erl [4]):
Table 1: SOA Patterns by Process Modeled
The SOA patterns used differ in the two modules. These
patterns were next analyzed in greater detail and we present
those results in the UML activity diagrams [14] below.
iii. UML Activity Diagrams
UML activity diagrams are a useful tool for analyzing
the flow of logic through processes [14]. The following
diagrams were created from the code in the application. The
B2B diagram has less workload complexity (due to the
absence of out-of-stock condition handling) than the store
diagram.
Figure 7: B2C UML Activity Diagram
Figure 8: Store UML Activity Diagram
C. Bare Metal Implementations
Two bare metal implementations were used in the
experiments. Both were configured to run the SPECjbb2015
suite in a Windows 10 environment. One run of the
SPECjbb2015 suite was configured to accept a workload of
90% online “B2C” and 10% store transactions. The other
workload was configured to accept the inverse (10% online
“B2C” 90% store). No other applications were running in
when these tests were executed and the RAM allocated to
the processes was 8GB.
D. Virtualization Hypervisor Architecture
The second environment used in the experiment used the
Microsoft Windows 10, Server 2012 R2 and the Hyper-V
hypervisor extension. The SPECjbb2015 software was
compiled inside the virtual machine (VM). A full physical
CPU, Network Interface Card (NIC) and all storage available
was allocated to the VM. In addition, RAM of 8GB was
allocated to make this VM very similar to the “bare metal”
configuration. An architecture diagram is shown below, ours
however uses only one VM rather than the two shown.
SOA Pattern B2C Store
Agnostic Context X X
Canonical Protocol X X
Domain Inventory X
Functional Decomposition X
Inventory End X
Redundant Implementation X
Service Data Replication X
start
Txi Sends SM an
OnlinePurchaseRequest
Select a Random
Customer
Retrieve Customers
Previous Purchase History
Reserve Specific
Quantity Of Each
Product
Calculate Total price
Add the Available
Discounts and
Coupouns
Customer Basket Validation
Max Products
Available
Many Products to
be Replenished
Send Requests to
"Alternate Super
market List"
Proceed to Check
out
Generate Reciept
Check Customer's Credit
Customer has
enough Credit
Customer doesn't
have enough Credit
Transaction Fails
Move the Purchased
Items from Store
Inventory
Debit the cost of Each
Item from Customer's
Account
Send Requests to
Alternate Super Market
List
Reserve the
Remaining Goods
Online
Purchasing
Activity
Delivery Options
Store Pickup
Door Delivery to the
Customer
Send the Reciept to the HQ
Stop
start
Txi Sends SM an
InStorePurchaseRequest
Select a Random
Customer
Retrieve Customers
Previous Purchase History
Reserve Specific
Quantity Of Each
Product
Calculate Total price
Add the Available
Discounts and
Coupouns
Customer Basket Validation
Max Products
Available
Many Products to
be Replenished
Throw an Exception
Proceed to Check
out
Generate Reciept
Check Customer's Credit
Customer has
enough Credit
Customer doesn't
have enough Credit
STOP
Transaction Fails
Move the Purchased
Items from Store
Inventory
Debit the cost of Each
Item from Customer's
Account
Send Suppliers a
Request, If any item runs
out from store
Send Receipt back
to HQ
Stop
Instore
Activity
Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 7
ISBN: 1-60132-458-8, CSREA Press ©](https://image.slidesharecdn.com/114227-250418134504-5ac877e8/85/Grid-Cloud-and-Cluster-Computing-and-Applications-1st-Edition-Hamid-R-Arabnia-19-320.jpg)
Grid Cloud and Cluster Computing and Applications 1st Edition Hamid R. Arabnia
- 1. Grid Cloud and Cluster Computing and Applications 1st Edition Hamid R. Arabnia pdf download https://ebookmeta.com/product/grid-cloud-and-cluster-computing- and-applications-1st-edition-hamid-r-arabnia/ Download more ebook from https://ebookmeta.com
- 2. We believe these products will be a great fit for you. Click the link to download now, or visit ebookmeta.com to discover even more! Internet Computing and Internet of Things 1st Edition Hamid R. Arabnia https://ebookmeta.com/product/internet-computing-and-internet-of- things-1st-edition-hamid-r-arabnia/ Bioinformatics and Computational Biology 1st Edition Hamid R. Arabnia https://ebookmeta.com/product/bioinformatics-and-computational- biology-1st-edition-hamid-r-arabnia/ Health Informatics and Medical Systems 1st Edition Hamid R. Arabnia https://ebookmeta.com/product/health-informatics-and-medical- systems-1st-edition-hamid-r-arabnia-2/ The Future of Health How Digital Technology Will Make Care Accessible Sustainable and Human 1st Edition Roberto Ascione https://ebookmeta.com/product/the-future-of-health-how-digital- technology-will-make-care-accessible-sustainable-and-human-1st- edition-roberto-ascione/
- 3. Divas Diamonds Death Danger Cove Cozy Mystery 15 Elizabeth Ashby https://ebookmeta.com/product/divas-diamonds-death-danger-cove- cozy-mystery-15-elizabeth-ashby/ Descartes and the ontology of everyday life First Edition. Edition Deborah J Brown And Calvin G Normore https://ebookmeta.com/product/descartes-and-the-ontology-of- everyday-life-first-edition-edition-deborah-j-brown-and-calvin-g- normore/ Powerful Understanding Helping Students Explore Question and Transform Their Thinking about Themselves and the World Around Them 1st Edition Adrienne Gear https://ebookmeta.com/product/powerful-understanding-helping- students-explore-question-and-transform-their-thinking-about- themselves-and-the-world-around-them-1st-edition-adrienne-gear/ Washington Manual of Medical Therapeutics Spiral 36th Edition Zachary Crees https://ebookmeta.com/product/washington-manual-of-medical- therapeutics-spiral-36th-edition-zachary-crees/ That Sik Luv Jescie Hall https://ebookmeta.com/product/that-sik-luv-jescie-hall/
- 4. Encyclopedia of Molecular Pharmacology, 3rd Edition Stefan Offermanns https://ebookmeta.com/product/encyclopedia-of-molecular- pharmacology-3rd-edition-stefan-offermanns/
- 5. Editors © CSREA Press PROCEEDINGS OF THE 2017 INTERNATIONAL CONFERENCE ON GRID, CLOUD, & CLUSTER COMPUTING Hamid R. Arabnia Fernando G. Tinetti CSCE’17 July 17-20, 2017 Las Vegas Nevada, USA americancse.org
- 6. Copyright and Reprint Permission Copying without a fee is permitted provided that the copies are not made or distributed for direct commercial advantage, and credit to source is given. Abstracting is permitted with credit to the source. Please contact the publisher for other copying, reprint, or republication permission. Copyright © 2017 CSREA Press ISBN: 1-60132-458-8 Printed in the United States of America This volume contains papers presented at The 2017 International Conference on Grid, Cloud, & Cluster Computing (GCC'17). Their inclusion in this publication does not necessarily constitute endorsements by editors or by the publisher.
- 7. Foreword It gives us great pleasure to introduce this collection of papers to be presented at the 2017 International Conference on Grid, Cloud, and Cluster Computing (GCC’17), July 17-20, 2017, at Monte Carlo Resort, Las Vegas, USA. An important mission of the World Congress in Computer Science, Computer Engineering, and Applied Computing, CSCE (a federated congress to which this conference is affiliated with) includes "Providing a unique platform for a diverse community of constituents composed of scholars, researchers, developers, educators, and practitioners. The Congress makes concerted effort to reach out to participants affiliated with diverse entities (such as: universities, institutions, corporations, government agencies, and research centers/labs) from all over the world. The congress also attempts to connect participants from institutions that have teaching as their main mission with those who are affiliated with institutions that have research as their main mission. The congress uses a quota system to achieve its institution and geography diversity objectives." By any definition of diversity, this congress is among the most diverse scientific meeting in USA. We are proud to report that this federated congress has authors and participants from 64 different nations representing variety of personal and scientific experiences that arise from differences in culture and values. As can be seen (see below), the program committee of this conference as well as the program committee of all other tracks of the federated congress are as diverse as its authors and participants. The program committee would like to thank all those who submitted papers for consideration. About 65% of the submissions were from outside the United States. Each submitted paper was peer-reviewed by two experts in the field for originality, significance, clarity, impact, and soundness. In cases of contradictory recommendations, a member of the conference program committee was charged to make the final decision; often, this involved seeking help from additional referees. In addition, papers whose authors included a member of the conference program committee were evaluated using the double-blinded review process. One exception to the above evaluation process was for papers that were submitted directly to chairs/organizers of pre-approved sessions/workshops; in these cases, the chairs/organizers were responsible for the evaluation of such submissions. The overall paper acceptance rate for regular papers was 23%; 14% of the remaining papers were accepted as poster papers (at the time of this writing, we had not yet received the acceptance rate for a couple of individual tracks.) We are very grateful to the many colleagues who offered their services in organizing the conference. In particular, we would like to thank the members of Program Committee of GCC’17, members of the congress Steering Committee, and members of the committees of federated congress tracks that have topics within the scope of GCC. Many individuals listed below, will be requested after the conference to provide their expertise and services for selecting papers for publication (extended versions) in journal special issues as well as for publication in a set of research books (to be prepared for publishers including: Springer, Elsevier, BMC journals, and others). • Prof. Nizar Al-Holou (Congress Steering Committee); Professor and Chair, Electrical and Computer Engineering Department; Vice Chair, IEEE/SEM-Computer Chapter; University of Detroit Mercy, Detroit, Michigan, USA • Prof. Hamid R. Arabnia (Congress Steering Committee); Graduate Program Director (PhD, MS, MAMS); The University of Georgia, USA; Editor-in-Chief, Journal of Supercomputing (Springer); Editor-in-Chief, Transactions of Computational Science & Computational Intelligence (Springer); Fellow, Center of Excellence in Terrorism, Resilience, Intelligence & Organized Crime Research (CENTRIC). • Prof. Dr. Juan-Vicente Capella-Hernandez; Universitat Politecnica de Valencia (UPV), Department of Computer Engineering (DISCA), Valencia, Spain • Prof. Kevin Daimi (Congress Steering Committee); Director, Computer Science and Software Engineering Programs, Department of Mathematics, Computer Science and Software Engineering, University of Detroit Mercy, Detroit, Michigan, USA
- 8. • Prof. Leonidas Deligiannidis (Congress Steering Committee); Department of Computer Information Systems, Wentworth Institute of Technology, Boston, Massachusetts, USA; Visiting Professor, MIT, USA • Prof. Mary Mehrnoosh Eshaghian-Wilner (Congress Steering Committee); Professor of Engineering Practice, University of Southern California, California, USA; Adjunct Professor, Electrical Engineering, University of California Los Angeles, Los Angeles (UCLA), California, USA • Prof. Louie Lolong Lacatan; Chairperson, Computer Engineerig Department, College of Engineering, Adamson University, Manila, Philippines; Senior Member, International Association of Computer Science and Information Technology (IACSIT), Singapore; Member, International Association of Online Engineering (IAOE), Austria • Prof. Hyo Jong Lee; Director, Center for Advanced Image and Information Technology, Division of Computer Science and Engineering, Chonbuk National University, South Korea • Dr. Ali Mostafaeipour; Industrial Engineering Department, Yazd University, Yazd, Iran • Dr. Houssem Eddine Nouri; Informatics Applied in Management, Institut Superieur de Gestion de Tunis, University of Tunis, Tunisia • Prof. Dr., Eng. Robert Ehimen Okonigene (Congress Steering Committee); Department of Electrical & Electronics Engineering, Faculty of Engineering and Technology, Ambrose Alli University, Edo State, Nigeria • Prof. Igor Schagaev; Director of ITACS Ltd, United Kingdom (formerly a Professor at London Metropolitan University, London, UK) • Ashu M. G. Solo (Publicity), Fellow of British Computer Society, Principal/R&D Engineer, Maverick Technologies America Inc. • Prof. Fernando G. Tinetti (Congress Steering Committee); School of CS, Universidad Nacional de La Plata, La Plata, Argentina; Co-editor, Journal of Computer Science and Technology (JCS&T). • Prof. Layne T. Watson (Congress Steering Committee); Fellow of IEEE; Fellow of The National Institute of Aerospace; Professor of Computer Science, Mathematics, and Aerospace and Ocean Engineering, Virginia Polytechnic Institute & State University, Blacksburg, Virginia, USA • Prof. Jane You (Congress Steering Committee); Associate Head, Department of Computing, The Hong Kong Polytechnic University, Kowloon, Hong Kong • Dr. Farhana H. Zulkernine; Coordinator of the Cognitive Science Program, School of Computing, Queen's University, Kingston, ON, Canada We would like to extend our appreciation to the referees, the members of the program committees of individual sessions, tracks, and workshops; their names do not appear in this document; they are listed on the web sites of individual tracks. As Sponsors-at-large, partners, and/or organizers each of the followings (separated by semicolons) provided help for at least one track of the Congress: Computer Science Research, Education, and Applications Press (CSREA); US Chapter of World Academy of Science; American Council on Science & Education & Federated Research Council (http://www.americancse.org/); HoIP, Health Without Boundaries, Healthcare over Internet Protocol, UK (http://www.hoip.eu); HoIP Telecom, UK (http://www.hoip-telecom.co.uk); and WABT, Human Health Medicine, UNESCO NGOs, Paris, France (http://www.thewabt.com/ ). In addition, a number of university faculty members and their staff (names appear on the cover of the set of proceedings), several publishers of computer science and computer engineering books and journals, chapters and/or task forces of computer science associations/organizations from 3 regions, and developers of high-performance machines and systems provided significant help in organizing the conference as well as providing some resources. We are grateful to them all. We express our gratitude to keynote, invited, and individual conference/tracks and tutorial speakers - the list of speakers appears on the conference web site. We would also like to thank the followings: UCMSS (Universal Conference Management Systems & Support, California, USA) for managing all aspects of the conference; Dr. Tim Field of APC for coordinating and managing the printing of the proceedings; and the staff of Monte Carlo Resort (Convention department) at Las Vegas for the professional service they
- 9. provided. Last but not least, we would like to thank the Co-Editors of GCC’17: Prof. Hamid R. Arabnia and Prof. Fernando G. Tinetti. We present the proceedings of GCC’17. Steering Committee, 2017 http://americancse.org/
- 11. Contents SESSION: CLOUD SERVICES AND COMPUTING Measuring SPECjbb2015 Virtualization Overhead and Application Workloads in a Microsoft Hyper-V Cloud 3 John Medellin, Frank Lee, Lokesh Budhi, Spandana Gennepally Using Smart Contracts and Blockchains to Support Consumer Trust Across Distributed Clouds 10 Stephen S. Kirkman, Richard Newman Design of an Architectural Framework for Providing Quality Cloud Services 17 Paromita Goswami, Sayantan Singha Roy, Ranjan Dasgupta Developing Enterprise Applications for Cloud: The Unicorn Application Framework 24 Marek Beranek, Vladimir Kovar, George Feuerlicht A Rapid Auto-Scaling Mechanism in Cloud Computing Environment 31 Chia-Wei Tseng, Ming-Shiun Tsai, Yao-Tsung Yang, Li-Der Chou EZPoll: a Progressive Implementation of Cloud-Based Polling Systems 35 Hans Dulimarta MCloud API For Managing Data in Multi-Cloud Environment 42 Maram Alhafdi, Ahmed Barnawi SESSION: DISTRIBUTED COMPUTING AND NOVEL APPLICATIONS + CLUSTER COMPUTING HPC Cluster Management with Open Source Software 49 Fernando G. Tinetti, Leopoldo J. Rios Group Mutual Exclusion Algorithm for Intersection Traffic Control of Autonomous Vehicle 55 SungHoon Park Park, BoKyoung Kim, YeongMok Kim Cost-Efficiency Comparison of an ARM Cluster & Intel Server 59 Jaden D. Weiss SESSION: POSTER PAPERS Design of a Tool for Analyzing Supercomputer Status using a Integrated Log Repository 65 Sung-Jun Kim, Jae-Kook Lee, Tae-Young Hong Correlation Analysis of OpenStack Log using Machine Learning Techniques 67 Ju-Won Park, Eunhye Kim
- 12. SESSION: LATE PAPERS - DISTRIBUTED AND CLOUD COMPUTING Evaluating Cloud Auto-Scaler Resource Allocation Planning Under High-Performance Computing Workloads 71 Kester Leochico, Eugene John A Review of Electric Vehicle Charging Impact on Power Grid 78 Azhar Ul-Haq, Saif ullah Awan
- 13. SESSION CLOUD SERVICES AND COMPUTING Chair(s) TBA Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 1 ISBN: 1-60132-458-8, CSREA Press ©
- 14. 2 Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | ISBN: 1-60132-458-8, CSREA Press ©
- 15. Measuring SPECjbb2015 Virtualization Overhead & Application Workloads in a Microsoft Hyper-V Cloud John M. Medellin*, Frank C. Lee, Lokesh Budhi, Spandana Gennepally Faculty and Graduate Assistants at the Master of Science in Information Systems Program University of Mary-Hardin Baylor Belton, TX 76513-2599, USA Abstract— Virtualized environments are a valuable component of most modern automation strategies. They allow for creation of logical systems that can be extended to accommodate a variety of needs in the computing community. The critical manager of this flexibility is the hypervisor software; it controls the direction and limitations of the resources available to execute workloads. Hypervisors are available in large varieties both from reputable vendors and the open source community. This paper executes experiments within the Microsoft Stack (Windows 10, Windows Server 2012 R2 and Windows Hyper-V 2016) and measures the incremental overhead added by the virtualization layers. The flexibility afforded by Hypervisors is not delivered without increase in computing cost. In addition to operating the internals of the Hypervisor software, there are also penalties to pay for virtualization of resources themselves. Virtualization overhead has been studied in a variety of formats but lately, it has focused on determining overhead driven by the impact of workloads on key components; data, network, memory or computation, and how different Hypervisors perform under certain types of applications. Our paper measures two types of workloads that are common in the retail industry; online shopping (business to consumer or “B2C”) vs shopping at the store. We use the SPECjbb2015 benchmark suite to configure different workload characteristics under each scenario and determine workload and overhead effects using the standard Microsoft stack. Our conclusions indicate that there are differences in overhead incurred by the virtualization layers and between the application workloads themselves. There seems to be an incremental combined overhead/workload penalty for more complex applications. The preliminary conclusion is that this virtualized environment gives better performance benchmarks to online shopping than to traditional store level sales based on the different complexities in the applications. Keywords— Hypervisors, B2B/B2C, Application Workloads, Retail Applications, SPEC Corporation, Microsoft Windows 10, Microsoft Windows Server 2012 R2, Microsoft Hyper-V 2016, Services Oriented Architecture, SOA I. INTRODUCTION Virtualized environments were first introduced in the 1970s as a means to extend processing cycle usage and avoid unacceptable levels of idle CPU resources. This concept has been further exploited in the past decade with the advent of Cloud environments. In the Cloud, we are able to model a variety of logical configurations that can help us achieve certain workloads depending on their characteristics [11]. We can adapt our architectures to take advantage of certain configurations that can lend better performance to different types of application complexity. A significant amount of work has been done in measuring the impact of workloads on the managers of Cloud environments. These managers (Virtual Machine Managers) are also called “Hypervisors”. Hypervisors are available both from traditional vendors (Microsoft in the Hyper-V product for example [5]) or on open source (like the OpenStack product [10]). These Hypervisors can lie on the bare-metal (right on the actual hardware) or as extensions to Network Operating systems as in Microsoft Hyper-V being an extension of Microsoft Windows Server 2012 Release 2. These configurations allow the deployment of “client” operating systems which are resident inside the virtual machines (for example a client-based Windows 10 implementation) and provide the execution environment for the applications themselves. Most of the recent body of knowledge has focused on the measurement of the various parts of the virtualization overhead using standard workload software (these are available suites that focus on producing a type of environment that is stable and has been used in research before). One of these suites is the SPECjbb2015 by www.spec.org. This standard workload generator can simulate a variety of transaction mixes and can enable very complex scenarios of business environments. The system simulates the transaction load of a supermarket chain and can be scaled to multiple supermarkets and multiple “central management” offices. The objective of this system is to measure throughput under an incremental volume scaling algorithm until the system is saturated; the system can no longer keep up with the load provided. Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 3 ISBN: 1-60132-458-8, CSREA Press ©
- 16. The fundamental objectives of this research are to measure the incremental overhead and workload added by virtualization on two common retail industry application patterns: online shopping and store shopping. We use the SPECjbb2015 to vary the transaction mix between online purchases and in-store purchases for each trial. We start by installing this environment on bare metal versions of Windows 10 and then follow by creating virtual machine images under Windows Server 2012 R2 with the Hyper-V 2016 hypervisor. The reduced throughput reported under each of these Cloud environments is due to the incremental overhead from virtualization. The two differences in workload execution architecture patterns are analyzed through use cases, software architecture patterns and UML activity diagrams derived from the SPECjbb2015 code. This document is structured as follows: • We present a summary of the related work that has been done in this area and how we have leveraged those conclusions on our research. • We structure an experiment which measures the impact of virtualization by first executing the SPECjbb2015 on bare metal under the Windows 10 operating system and then proceed to implement that environment under Windows Server 2012 R2 Network OS with the Hyper-V 2016 Hypervisor. The incremental layer of abstraction is measured and reported by the throughput achieved in the SPECjbb2015 suite. • We alter the transaction load of SPECjbb2015 to create one scenario with Web Online Shopping (“B2C”) transactions and one other with Store (“brick and mortar”; “B2M”) sales transactions. The differences in application throughput performance are also reported for each scenario. This initial research is the first step in understanding the effects of various types of application design patterns for Cloud applications. We believe the differences can be explained (and in further research optimized) by better utilization of standard design patterns in Cloud-based applications. II. RELATED WORK Extensive work has been done on the effects of virtualized environments and some of the concepts related to creating virtual systems date back to the 1980s [13]. The full virtual cloud concept was only possible after the creation of software-defined systems that managed servers, storage, network, CPU and other resources known as Virtual Machine Monitors or more commonly as hypervisors [3]. The hypervisor establishes policies for sourcing of resources to the various processes that are resident on a particular infrastructure. Hypervisors are usually complemented by other software to create a Cloud Management Platforms (CMP) that enhance management of the virtualized environment [6]. A. Cloud Framework and Hypervisors Freet, Agrawal, Walker and Badr [6] provide a review of the general characteristics of the CMP that make hypervisors effective. They reviewed Eucalyptus, OpenStack, CloudStack, OpenNebula, Nimbus and Proxmox (all open source CMPs). Their analysis concludes that OpenStack and CloudStack have over 30 times more messages that some of their other competitors (meaning they are more top of mind in the development community). They proceed to review the frameworks in relation to Xen, KVM, Virtual Box and ESX/VMware and requirements for data center virtualization and infrastructure provision. In that study, various types of workloads are varied through each candidate hypervisor and the throughput is reported. We have adopted a similar workload variation approach and report throughput statistics in our methodology. B. Virtualization Overhead As mentioned above, virtualization of resources can be a costly proposition. Oi and Nakajima [9] explored the effects of performance loads on the Xen environment vs others. They determined that the performance of Xen could be enhanced in a virtualized environment by adjusting cache sizes in some applications. Although Xen was never able to provide the throughput that Linux could, (both operating on the hardware without other Operating Systems present) it was shown how certain configuration changes could be effected to drive higher throughput in the Xen hypervisor environment. The point of varying configurations impact on workload efficiency is conclusively illustrated in that work. The benchmarking system used was SPECjbb2001 and the effects of Network Interface Cards (NIC) were isolated so the workload could be measured in memory usage and throughput. We have also adopted the SPEC performance suite but with an application development focus on the results. Al Jabry, Liu, Zhu and Panneerselvam [1] studied the effects of hypervisor overhead along the following resource usage: disk i/o, CPU, memory and VMM (hypervisor). Their tests were conducted using VMware, Virtual Box (Oracle Corporation) and Windows Virtual PC (Microsoft). Their work benchmarked a standard load in each hypervisor environment and used IOzone to quantify load on disk i/o, RAMSpeed to quantify the impact on memory and UnixBench, to indicate the effect on CPU. The residual resource usage is attributed to the VMM (hypervisor) resource. In their conclusion, they note that hypervisor overhead is considerably low on the memory and disk i/o segments while it is much higher on the CPU usage. This 4 Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | ISBN: 1-60132-458-8, CSREA Press ©
- 17. study points to the need to choose hypervisors in relation to the application workload needs and points to the Microsoft solution as being the more balanced of the hypervisors reviewed with those workloads. We incorporated the usage of the Microsoft stack into our experiments because of the tight integration within itself. Chen, Patel, Shen and Zhou [2] studied virtualization overhead across multiple VMs running under Xen in cloud environments. They also found that the larger resource usage was attributable to the CPU. They also propose a series of equations that are remarkably accurate in predicting the lateral scaling of workloads on all components based on the observed results of the application under study. Their recommendation for future work is to vary the application characteristics (workload profiles) and further fine-tune their predictive analysis. We incorporate those recommendations by varying our workloads based on B2C and store transaction loads. C. Application Workload Research Based on the research referenced, there is a significant impact on utilization of CPU from the overhead generated by the hypervisor. Further the impact is based on the type of application that is operating in the virtualized environment. NasiriGerdeh, Hosseini, RahimiZadeh and AnaLoui [8] measured throughput degradation on Web applications using the Faban suite (a web-based workload generator). They simulated the behavior of heavy transactional Web applications that tend to be very network intensive. Their work also measured the effect on memory, disk i/o and CPU. They concluded that a disproportionate difference exists in CPU resources due to the translation of domain addresses. This work further confirms that the principal resource difference is the CPU utilization even when workloads may be more i/o bound (the penalties associated where in finding addresses; a CPU task, not access to the actual addresses in the Web environment; an i/o task). We incorporate this research by focusing on actual compute power utilization rather than network or disk access. The SPECjbb2015 suite is focused on exhausting the compute resources rather than the disk (i/o) or network resources. San Wariya, Nair and Shiwani [12] focused their research on benchmarking three hypervisors; Windows Hyper-V, VMWare/ESXi and Citrix Xen in three cloud games; 3D Mark 11, Unigine Heaven and Halo. The objectives of their study are to identify which hypervisor was better from a cloud gaming workload perspective. The three performed differently in each category but were mostly lead by the VMWare product. For our purposes however, the HALO benchmark (number of frames per second) is probably the most predictive of workloads that are CPU bound. In this category, Hyper-V performed 7% ahead of VMWare and 57% ahead of Citrix Xen. This was another reason for selection of Hyper-V as the hypervisor for our test suite. D. SOA Application Design Patterns Standardized software architecture patterns provide a series of guidelines for developing functionality that is common across a variety of applications [14]. The work in this domain has been futher elaborated in Services Oriented Architecture (SOA) patterns that can be used to overcome certain design dilemmas. Thomas Erl (with foreword by Grady Booch) [4] presents a comprehensive set of architecture patterns that are common in business. These patterns were matched to the two use cases (B2C & store below) and confirmed in the source code of SPECjbb2015. Workloads can be modeled in a variety of ways in the application architecture and each can achieve their goals in a different way. If we however, standardize those patters with relation to use cases, there is a greater possiblity for the results to be extrapolated by other application researchers in this field. This approach gives a better chance of replication of the results for other studies relying on our work. A frequently used SOA application design pattern for abstracting logic from traditional application design nomenclature is shown in Figure 1 below [4]. Other SOA design patterns were confirmed in the implementation of the use cases below in the application code itself. Figure 1: from traditional design pattern to services architecture design pattern (Erl et. al.) E. The SPEC Benchmarking Suite The SPECjbb2015 (Figure 4) constitutes a Business to Consumer and Store to Consumer simulation for a Supermarket Chain. The model can be extended to include several supermarkets and several central offices in a variety of virtual machine settings. The benchmarking suite can be configured in a variety of business transaction settings so that different business patterns can be simulated (e.g., web sales versus physical store sales). The system is owned and licensed by spec.org [15] which is a consortium of major IT companies that have agreed on a set of principles to guide the performance benchmarking process. The system progressively injects transaction loads into the environment until saturation is reached. A sample output of these results is seen in Figure 3. In that graphic the system begins to stress at around the 5,200 java Operations Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 5 ISBN: 1-60132-458-8, CSREA Press ©
- 18. Per Second (jOPS) with a range of 5K (median tolerance) to 50K (max tolerance). The system reaches saturation (min tolerance) at around 6,700 jOPS and 60K. We report our results using the total transactions up to saturation. Figure 4 is a graphic representation of the architecture of the system. Figure 3: Sample SPECjbb2015 Benchmark Output www.spec.org Figure 4: SPECjbb2015 Architecture www.spec.org The SPECjbb products have been in existence since the late 1990s and are useful because of their industry acceptance. For example, Karlsson, Moore, Hagersten and Wood [7] used an earlier version (SPECjbb2001) and another application benchmark (ECPerf) to differentiate effects of cache misses between different types of applications. III. EXPERIMENT DESIGN As discussed above, experiments were designed where the same application (SPECjbb2015) was installed on: a) Bare metal with Windows 10 b) VMachine: Windows Server/Hyper-V/Windows 10 Two transaction variations were used; one increased the B2C transaction types to be 90% of the sales while the second one increased the store sales to 90% of the sales. B. Application Architecture Patterns The application patterns were analyzed by matching use cases, identifying SOA design patterns and preparing activity diagrams for each version of the application configured. The details are reported below. i. Use Case Analysis Use cases are a functional decomposition tool that illustrate the process interactions between actors in applications [14]. The processes that we have selected in the SPECjbb2015 suite are fairly standard and follow similar patterns. The use case diagram for the store architecture is similar to this one (www.UML-diagrams.org), the “adornments” in the graphic describe the usage of artifacts. Figure 5: Store Use Case Diagram The use case diagram for B2C sales architecture pattern is similar to the one below (www.UML-diagrams.org) : Figure 6: B2C Use Case Diagram At a high level, both use cases are similar in providing functionality for the user to execute a purchase. However, when inspected the code we found that different design patterns and workloads were implemented. Most of these differences are due to the inventory on-hand function at the physical store. If the system detects an out-of-stock condition, it will proceed to cancel and back out the transaction. This process will take more resources than fulfilling the order from other stores (as it does in the B2C scenario) because it has to place the order items into inventory and invalidate the order itself (error exception in the store UML activity diagram below). ii. SOA Design Patterns in SPECjbb2015 We inspected the code again to determine which patterns were evident. Even though the processes are very similar 6 Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | ISBN: 1-60132-458-8, CSREA Press ©
- 19. from a functional perspective, they are less similar from a pattern design perspective. The SOA design patterns that were used (in our opinion) are as follows (from Erl [4]): Table 1: SOA Patterns by Process Modeled The SOA patterns used differ in the two modules. These patterns were next analyzed in greater detail and we present those results in the UML activity diagrams [14] below. iii. UML Activity Diagrams UML activity diagrams are a useful tool for analyzing the flow of logic through processes [14]. The following diagrams were created from the code in the application. The B2B diagram has less workload complexity (due to the absence of out-of-stock condition handling) than the store diagram. Figure 7: B2C UML Activity Diagram Figure 8: Store UML Activity Diagram C. Bare Metal Implementations Two bare metal implementations were used in the experiments. Both were configured to run the SPECjbb2015 suite in a Windows 10 environment. One run of the SPECjbb2015 suite was configured to accept a workload of 90% online “B2C” and 10% store transactions. The other workload was configured to accept the inverse (10% online “B2C” 90% store). No other applications were running in when these tests were executed and the RAM allocated to the processes was 8GB. D. Virtualization Hypervisor Architecture The second environment used in the experiment used the Microsoft Windows 10, Server 2012 R2 and the Hyper-V hypervisor extension. The SPECjbb2015 software was compiled inside the virtual machine (VM). A full physical CPU, Network Interface Card (NIC) and all storage available was allocated to the VM. In addition, RAM of 8GB was allocated to make this VM very similar to the “bare metal” configuration. An architecture diagram is shown below, ours however uses only one VM rather than the two shown. SOA Pattern B2C Store Agnostic Context X X Canonical Protocol X X Domain Inventory X Functional Decomposition X Inventory End X Redundant Implementation X Service Data Replication X start Txi Sends SM an OnlinePurchaseRequest Select a Random Customer Retrieve Customers Previous Purchase History Reserve Specific Quantity Of Each Product Calculate Total price Add the Available Discounts and Coupouns Customer Basket Validation Max Products Available Many Products to be Replenished Send Requests to "Alternate Super market List" Proceed to Check out Generate Reciept Check Customer's Credit Customer has enough Credit Customer doesn't have enough Credit Transaction Fails Move the Purchased Items from Store Inventory Debit the cost of Each Item from Customer's Account Send Requests to Alternate Super Market List Reserve the Remaining Goods Online Purchasing Activity Delivery Options Store Pickup Door Delivery to the Customer Send the Reciept to the HQ Stop start Txi Sends SM an InStorePurchaseRequest Select a Random Customer Retrieve Customers Previous Purchase History Reserve Specific Quantity Of Each Product Calculate Total price Add the Available Discounts and Coupouns Customer Basket Validation Max Products Available Many Products to be Replenished Throw an Exception Proceed to Check out Generate Reciept Check Customer's Credit Customer has enough Credit Customer doesn't have enough Credit STOP Transaction Fails Move the Purchased Items from Store Inventory Debit the cost of Each Item from Customer's Account Send Suppliers a Request, If any item runs out from store Send Receipt back to HQ Stop Instore Activity Int'l Conf. Grid, Cloud, & Cluster Computing | GCC'17 | 7 ISBN: 1-60132-458-8, CSREA Press ©
- 20. Exploring the Variety of Random Documents with Different Content
- 21. life be taken from me," and Corman ceased rowing, and gazed ruefully towards Boseham. "Oh, Corman, look! the man is galloping along the shore, and ——why, there are several more men coming down. What are they going to do?" Corman and Ædric remained for a few moments in speechless curiosity. The man on horseback had galloped furiously up to the men, and was gesticulating rapidly. The men dispersed and ran about the shore. At last they all seemed to be running to one spot. They all collected round something, the man on horseback appearing to be energetically directing them. "Why, it's one of the Boseham boats they are launching, I do believe," said Corman. "So it is, and now they are getting into it. What do you think it means?" "I think they are going to row after us." So saying, Corman began rowing again as hard as he could. They were about three-quarters of a mile away, and had to go about a mile more before they could reach the "hard," or landing place, on the other side of the creek, for they had to row out of the little creek, at the head of which Boseham stands, and cross the larger creek that wound its muddy way up to within a mile of Cissanceaster. The tide was running out strongly, and this was all in their favour, for as they got farther down the stream ran stronger. Corman knew the importance of making the full use of the tide, and he strained every muscle to get into the main channel.
- 22. The other boat was now manned, and the crew were rowing vigorously, but unscientifically. The horseman had got in, and was steering. "They are not gaining much, if at all," said Ædric. Corman said nothing. He had need of all his strength and breath; the drops of perspiration on his brow told how hard he was working. The clumsy raft went sluggishly along in spite of all his toil, and the other boat came nearer. "Why can't I row? I know how; I have often done it at home. I could at least take one oar." Corman shook his head, and rowed hard. Nothing more was said by Ædric, and the oars splashed and the water gurgled under the unwieldy logs of the raft, as it slushed its way through the water. "They are gaining on us now. Well done!" cried Ædric, as one of the men twisted his oar under the water, and was knocked by the handle of it against the next man, and so into the bottom of the boat. It was comic to see his legs go up in the air, and to hear the shouts of wrath from the helmsman.
- 23. How Corman and Ædric fled before Berchthune "That has stopped them a bit. Now if they only would stick in the mud! the tide is falling fast, and they couldn't get off." Corman was getting tired with his exertions. It was quite clear, unless some accident happened to the other boat, they must be caught. They were so near to each other now that Ædric could distinguish the men. They were all strangers to him. The man who commanded was a tall, grey-bearded man, muscular and wiry. He wore a helmet and linked mail shirt, across which a chain hung supporting a two-edged battle axe; his keen eyes glared from under
- 24. thick, bushy, grey eyebrows, and two wings of a hawk attached to his helmet gave him a very war-like air. "Who can he be, I wonder?" said Ædric. Corman only shook his head by way of answer, and kept rowing desperately, but there was evidently no chance. Suddenly an idea struck him. "Ædric," he gasped, "dost thou see any shallow spot ahead over which we could go, but on which they would stick? If thou dost, point to it, and I will row over it." Ædric looked about; the sea was so muddy that it was difficult to tell where the deep water was, but the current ran in stronger eddies, and with more of ruffle on its surface in the channel, and the boy saw one bank that he thought would do. They had now got to the part of the creek where the Boseham lake, or creek, joined the arm that went up to Cissanceaster. There was a long spit of mud running out from the western shore. If they could pass over this they would gain a good bit on their pursuers, who might, perhaps, be tempted to follow them, in which case they would inevitably run aground, and would have to remain for some hours. "I see a lane of deeper water across that bank there, only you must row very hard for it. It is some way off yet!" cried Ædric. Corman tugged at the oars, the awkward raft moved hardly any quicker, and the drops of perspiration rolled off the monk. Nearer and nearer the other boat came after them. The steersman was laughing. Ædric could see his great mouth opening in a broad grin of triumph. The men were not rowing nearly so hard now, and he could hear them talking. They were quite confident of success.
- 25. "Pull, Corman, pull! we are just going into the shallow part." And the poor monk rowed harder than ever. His eyes were straining and bloodshot, and the muscles of his neck stood out like knotted cords. The bow of the other boat was only a few yards off. The man in the bows had put his oar in, and was standing ready to jump on board the raft. The water curled under the bows. Suddenly the man in the bows was jerked violently forward, a large rush of water spread over the yellow surface ahead, and a wild shout of joy rang out from Ædric. "They are ashore, they are ashore! Hurrah!" And so it was: the boat, drawing quite two feet of water, had plunged into the mud, and was now stuck fast. All was instantly confusion and clamour on board. The chieftain stormed and raged, notwithstanding it was entirely his own fault; for he had not followed the wake of the raft, but had tried to cut it off. The raft was still in comparatively deep water, and was going away merrily. The men on board all stood up, and pushed and tugged at their oars, but as fast as they pushed their oars in, and moved the boat at all, they pulled her on again by trying to get their oars out of the deep, clinging, holding mud. Fierce imprecations and abusive epithets flew from the commander, but all to no purpose. "Get out, men! out with ye, or we shall remain here for ever. See how the tide is falling!" shouted the old man. The men tumbled over the gunwale into the shallow water, but they could hardly have done a more useless thing. Instead of pushing the boat off they only pulled it all the deeper into the mud; for not being able to obtain any foothold, they hung on to the sides of the boat to prevent themselves sinking in. It was a ludicrous sight
- 26. to see all these strong men hanging round the boat, wallowing and plunging in the black, clinging mud. The helmsman grew more and more furious, the more it became apparent that their position was hopeless. The men, disgusted with the mud and their fruitless exertions, tried to get in again, and the sight was still more comic, as they all struggled to climb over the side of the high and awkward boat. Their muddy legs all had the appearance of wearing long black silk stockings, and as they wriggled and plunged, they gradually became covered with the same horrible, greasy, shiny coating. Sometimes a man would be seen to raise himself up, get one leg over the gunwale, lie down on his side, and try to roll himself into the boat, his other leg would wave in the air, and just as he was succeeding, some of the other men, intent on their own endeavours, would pull the boat too much down on that side, and he would roll over into the mud again. At last one or two succeeded in getting in, and the others, with their assistance, were hauled over the side, not without much bruising of legs and arms, and a plentiful bedaubment of mud. Meanwhile Ædric and Corman were getting on well. The monk had rested a little when he saw that they had got far enough away to be safe from any arrow, supposing the men had bows and arrows with them, and he and Ædric were laughing at the miserable plight of their pursuers. Suddenly Corman began to row vigorously again. He had looked round, and instantly worked as hard, or harder, than ever at the oars. "Why, Corman, what is the matter?" said Ædric; but the monk did not answer. Ædric looked about, puzzled; there was no other
- 27. boat in sight, and the men were still far too busy trying to get into their boat to be thinking of any means of pursuing them, even if they had a chance of finding any. But while Ædric was wondering what had caused these renewed exertions of brother Corman, the raft came to a stop. It also had run on the mud. Their position was now singular, and very tantalising to both parties, but especially so for Corman and Ædric, for a few strokes more or a few inches more water and they would have been over the bank and into the little lake that ran into the deep channel on the other side. But there was no help for it. They could not push the logs of wood across, tied together as they were, and they were compelled to sit patiently and watch the struggles of the men in the other boat. These latter had at last got in again, and a loud shout told Corman and Ædric they had discovered that they also were aground. "What shall we do?" asked Ædric, ruefully. "Sit here, my son, until the Lord sends the water back again." Poor Corman was not sorry altogether. It had been a terrible trial of his strength, and he had pluckily answered to it; but he was very exhausted. Fortunately he had the pitcher of water on board, which he had put there in case Ædric should want any, or feel faint, and it now came in very usefully. After taking a long draught, he uttered a sigh of satisfaction, and stretched himself at full length on the raft, closing his eyes and folding his hands together on his chest. Ædric pushed a skin under his head, but the monk took no notice. The boy would have liked to have talked, but he respected Corman's fatigue, and watched the other boat's crew instead. They were doing nothing, sitting listlessly on the sides of the boat, some
- 28. with their black legs hanging over, some with their legs inside, all looking disconsolate and foolish. They evidently had no bows with them, or they would have tried a shot at the raft. The tide had now gone down a long way, and both boat and raft were left high and dry. Corman still slept, and Ædric was beginning to be very weary of their position, when he thought he heard some one hailing them. He looked about, but could see no one. Thinking it was his fancy, he was going to lie down when again he heard a voice calling, and this time there was no doubt it was some one calling Corman. The boy instantly awoke the monk, who sat up and rubbed his eyes with a dazed look. "Corman, there is someone calling you." "Is there? Where?" said Corman, sleepily. "I don't know where. Listen, there it is again." Corman got up and stood upon the raft, which had by this time settled down with its weight into the mud. He looked about; the tide had got down so low that the mud banks in places obscured a view of the water. But as Corman looked round he caught sight of a small boat in the Cissanceaster channel as near to him as it could get, which was about a quarter of a mile off however. As soon as the men—for there were two—in the boat saw Corman, they shouted to him again. "Hullo!" cried Corman; "what dost thou want?" A confused collection of sounds answered. "I can't hear thee," shouted Corman. "Who art thou?" "Wevcumfrolfrid" was all Ædric could make out. "What does that mean?" said Corman. "Speak more clearly," he shouted.
- 29. Again the incomprehensible sound came back. "Well, they've got very weak voices, whoever they are," said Corman. "We've come from Wilfrid," came at last distinctly across the mud. "They have come from Wilfrid," cried Ædric, joyously. "We shall escape, then, after all." "I don't know that," said Corman. "How are we to get to them, or they to us?" At last an idea occurred to him. He got up again. "Hast thou any mud-pattens?" he shouted. No answer. He yelled out his question again. This time the word "No" reached him. "Canst thou not get any?" he yelled. "We'll gongetsome." "Well, that is sensible," said Corman, as he saw the boat go off towards the opposite shore. CHAPTER XV. "THE CRUEL CRAWLING FOAM, THE CRUEL HUNGRY FOAM." "Well, Ædric, if we can once get over to the other shore we shall be all safe, for Wilfrid is feared by all these South Saxons in a way that I never could understand." "But who dost thou think they are who are pursuing us?" "It must be the Eorldoman Berchthune."
- 30. Corman had now stretched himself out again, and was preparing to have his doze out. Fortunately, the weather was fine. Their situation was uncomfortable enough with fair weather; it would have been deplorable had it rained. The little raft lay stranded on a wide-stretching bank of mud; all round little rivulets washed their muddy courses out of the soft ooze. On one side, but at some distance, a belt of shingle, marked with a long brown streak, the boundary of the sea at high water, was surmounted by a brown growth of wind-blown bushes, relieved here and there by a weird oak-tree, whose blighted growth appealed in outstretched leafless branches to the north-east to protect it from the violent treatment it always received at the hands of its tormentor, the south-west wind; above, a grey sky, windless and still, while all the world below looked sodden, and muddy, and brown. On this world of mud a sea-gull or two were having an eager feast, not unaccompanied by an occasional fight over some succulent crab or juicy winkle, while a curlew dipped its curved beak among the brown sludge, or plaintively cried to its more fortunate mate. Overhead a heron winged its way, looking sardonically down on the dot of the raft and the somewhat larger speck of the boat. It was a dull, dreary scene— a world of mud, a world of wood, a world of grey and brown. Ædric looked at it all wearily enough. He began to feel sleepy too. It seemed so odd to be so close to their enemies, doing nothing, and yet perfectly safe. They were not more than five hundred yards off, and in the perfect quiet he could hear the voices of the men as they occasionally spoke. Gradually he dozed off. The seagulls came nearer, the crabs crawled up on to the edges of the raft, and the lobworms busily
- 31. raised their piles all round. So passed an hour. But what is it that causes the crabs to sidle away, and the gulls to get up on circling wings, screaming the while? "Wake up, Corman, wake up, Ædric, and see what your pursuers are doing," the wild birds seemed to cry. Weary of doing nothing, the idea had occurred to Berchthune to make a movable kind of platform of planks, by which two men could approach the raft. By laying down one set of boards and then standing on them, they were able to lay another set ahead, then getting on these, they were able to pull up the others, and slide them past and place them ahead again, and so they were able to make laborious but sure way up to their prey. In this way they had already advanced about fifty yards, and were getting more adroit in moving the boards. Heavily Corman was sleeping, and Ædric was far away in dreamland. Nearer and nearer the boards were being pushed; not without much noise and mirth from those in the boat, however. Several times the two adventurous ones had, in the confidence of their skill, gone too much to the side of their treacherous platform, with the result that they had slipped into the fathomless mud, and had to crawl ignominiously back upon their fickle plank, blacker and humbler men. Each of these checks to their pride had evoked shouts of laughter from their comrades and showers of abuse from Berchthune, who was fretting at the delay. Ædric was dreaming blissfully, and Corman still snored. Nearer and nearer the men approached, when a shout from their comrades urged them to more activity. The other boat had been seen returning from the Selsea shore. It ran on the mud at the
- 32. nearest point to the raft, and a man was seen to get out and walk over the slippery surface towards Corman and Ædric. "Why, he's got boards on his feet!" said the begrimed and weary South Saxon, as he squatted on his precarious plank to look at the strange spectacle, disgusted at the mean advantage of the other man. Quite safely the man slithered his way over the mud, carrying four flat boards in his hands. He had already gone nearly half the distance, and this in about five minutes; while the enterprising South Saxons had taken nearly an hour to get over an equal space. "Get on with ye, sluggards, or they will escape yet!" shouted Berchthune, stamping with rage at the idea of his game getting away, after all the hours of waiting on the mud, and the certainty of its falling into his hands at last, on which he had consolingly counted. The two South Saxons now realised that they must make the most desperate exertions if they hoped to get to the raft before the other man. They tugged at their boards—splash they went, into the mud ahead; quickly they got upon them—splash came the last ones they had trodden on out of the mud behind; they toiled at them to put them into their places, then jumped upon them, and once more heaved at their last resting-place. They had no time to look up, splash—slosh—heated work; grimy, filthy, slimy toil—and all the time the crew shouted to them, cheering them on, and encouraging them to fresh exertions. Brother Corman was well avenged for the trouble they had given him in the morning. The men were a great deal nearer the raft than the other man was; but he was going on steadily, and well. And in spite of all the South Saxons could do, the boards would stick in the mud, and their labour was
- 33. terrific. Their plight was piteous: the perspiration rained off their foreheads, and formed little lanes of white down their muddy faces. And all the time Berchthune yelled at them, and the crew hied them on. And now the men were not more than ten yards distant, while the other man was about the same. The excitement on board the boat became intense, for their men, going as they were in a line from them, seemed to be much nearer than the other man, whose whole distance was visible. "Make a jump for it!" roared Berchthune. "By Woden's beard, I'll have ye flayed alive if ye don't beat that 'nithing' there." The men tugged amain, but, alas for their success! they could not get their last resting-place up; they had, in their eagerness, placed the board they were standing on too far away from the one they had just left. They leant over the mud, they stretched themselves, they gasped, they dripped, but all to no purpose, and, worse than all, their last standing place began slowly to increase its distance. They had placed their boards on the slippery brow of one of the many little rivulets which drained the mud-banks, and as they leant over to get at the other planks left behind, all their weight, being on one side, caused the boards to lift at the other end, and begin slowly to slide down into the little gully. One of the men had reached over so far that, as the board receded, he fell forward on his face in the mud, clutching desperately to the other planks. The other man was just able to recover his balance before too late. "Hold on to my legs, man, can't thee?" roared the prostrate South Saxon, as loud as he was able, for his mouth was very near
- 34. the mud. The other man did as he was told. The situation was now too ludicrous, even for the man who was hastening, as fast as his awkward mud-pattens would allow him, to rescue Corman and Ædric. He stopped still and begun to roar with laughter. By this time Corman was beginning to be aware that there were other existences besides his own. He sat up, rubbed his eyes, looked about him, and could scarcely take in the situation. When he did he also burst out laughing, and Ædric, waking up, was astonished to see Corman sitting on the raft, his mouth wide open, and peals of laughter shaking him from head to foot. The unfortunate South Saxons were not nearly so much amused; the wretched one, who was now acting as a kind of animated tow-rope to the other planks, was hanging on grimly to the tenacious boards, while his comrade held on fast to his ankles and all the time the other boards were slowly slipping over the ooze. Neither man dare let go, and yet there was no hope of being able to pull the obstinate boards out of the mud, as there was no purchase by which they could be raised, and they were besides slimy with mud. For a minute the tension lasted; then slowly the man's hands slipped off the greasy planks, and he lay spread out, face downwards, on the ooze. The other South Saxon still held on to his legs, and the two, now that his comrade had let go of the firmly- imbedded planks, glided more speedily into the bed of the little rivulet. There was no danger of the prostrate man sinking into the mud provided he did not attempt to walk. The long weed-like grass that spread over the surface kept him up, so long as he lay outstretched; but he wanted to get on the boards on which his
- 35. comrade was seated, and the difficulty was how to do it. He wriggled and twisted, and sank his knees into the slime, but at last he succeeded in rolling himself down sideways on to the plank; and there the two men sat, disconsolate and helpless, within six yards of Corman and Ædric. All this time the Eorldoman Berchthune was shouting himself hoarse with abuse at the wretched adventurers, and Corman and Ædric were enjoying the sport. Their rescuer had now waddled up to them. Corman knew how to use mud-pattens, but the difficulty was how to carry Ædric. He could use one leg, and they managed by putting one mud-patten on his foot, and holding him between them, to get him off the raft. The South Saxons, seeing their prey escaping them, when they had so nearly grasped it, and urged on by the abuse of Berchthune, determined to make one more effort. Profiting by his experience of the buoyant nature of the mud, if only its properties were clearly understood, the South Saxon who had wriggled on to the planks beside his comrade determined to try the plan again. It was only six yards—only three times his own length—and the mud-pattens were not yet adjusted. Throwing himself forward on to the mud, he began to wriggle over it towards the raft. The other man, not to be out- done, began doing the same. "Quick, Ædric, or we shall be too late after all," cried Corman. The South Saxons were just reaching the raft as Corman and their deliverer assisted Ædric off between them. Wildly their pursuers flung themselves upon it. The others were only a few paces off. Without hesitating a moment, the first South Saxon reared himself erect on the raft, and sprang fiercely after the retreating
- 36. figures. He just managed to reach the skirts of Corman's frock, and plunged knee-deep in the mud. He held on to the poor, old worn gown of the monk, who struggled to wrench it out of his grasp, while Ædric and the other man pulled at Corman. Suddenly there was a crack, and the torn handful of Corman's garment remained in the South Saxon's hand, who sank deeper in the yielding mud with the recoil. The other South Saxon had been more prudent; he stood upon the raft and looked at the now secure Corman and Ædric, and at his miserable comrade—for miserable he was, far more so than at first sight appeared. He wriggled and struggled to get out; plunged his hands and arms up to their elbows in the mud. The more he strove, the more hopeless his position became. Deeper—deeper, down he sank—the mud was now up to his waist. If only he could get one leg out, or throw himself flat upon the mud again; but the suction of the mud was upon him. Its awful grasp had got sure hold of him. "For the love of Valhalla, lend me a hand!" he roared. "I can't, man. I can't reach thee!" cried the other. "Give me that oar—give me them both. Quick!" The oars were flung to him; he placed them under his armpits, and so low had he sunk that he rested on them. For a time they bore him up, but the slight sticks, only roughly flattened at the end, began to sink too; and the pain in his shoulders was acute. His situation was desperate, for although he was being only very slowly engulfed now, yet none the less was the progress very sure. The tide had begun to rise—it was coming in rapidly. Would there be time for the raft to float before he was suffocated, or would the sea
- 37. flow over his head first before there was water enough to float it? It was a desperate hope. Meanwhile, Corman and Ædric were safe in the boat Wilfrid had sent for them, and were far away on the other side.
- 38. How ye South Saxon was held by ye mud, and naught could save him:
- 39. Up and up flowed the tide. The sea gulls had had their feast of crabs, and were screaming overhead. The wretched man's eyeballs were starting from his head; his head was sunk between his shoulders. Up and up crept the tide. The lobworms had ceased to pile their little heaps; the crabs were playfully scampering to meet the crawling froth, pushed further and further with each succeeding wavelet. No hope! the water has reached his chin; the slimy froth and scum of the mud forms a collar round the doomed man's neck. One more prodigious effort, one despairing, gasping heave. No good! The hands are clasped over the mouth, with the instinct of self- preservation, even in inevitable death; but the water knows no barrier. The froth bubbles up, it is on a level with the lower lip, each wave and ripple washes higher, now the mouth is covered. With a desperate wrench, the gasping man raises his mouth above the water, but, unable to keep up the strain, his head sinks again, and this time the cruel water has reached the nose. The head falls down, a few bubbles, a little brown patch, hardly to be distinguished from seaweed, around which the yellow froth laps in the ripple, is all that marks where a strong man has died. Soon even that will have disappeared, and the place that knew him shall know him no more. The sea had been washing round the raft for some minutes, but the water-soaked logs were heavy, and had been sucked into the mud. The drowned man's head had been entirely covered before the awkward structure showed any signs of lifting. Indeed, the water was nearly floating over it, and the South Saxon had begun to dread a similar death to that of his comrade, when the raft gave a lurch, and once more was afloat. The man had no oars, or anything to
- 40. propel it with; but as the other boat would be afloat also before many minutes, they would come and pick him up. Presently the idea occurred to him to push the raft with one leg on the bottom; in this way, and with a favouring tide he was enabled at last to reach his companions. The Eorldoman Berchthune was very sullen, and greeted the man with violent abuse for not having made more haste at first; and this was all the misguided ceorl got for having volunteered on a perilous enterprise: for having been face to face with death, and that almost the slowest, most lingering, which could happen to man. But then in those days what were men made for but for death? The tide had now risen high enough to float the boat. Berchthune was about to give orders to shove her off the bank, when a horseman galloping hastily down the shingle on the shore, and riding his horse as far out as he dared, shouted to the boat: "Cædwalla has been made king of Wessex, and is marching upon us." There was now no thought of pursuing Ædric. Orders were instantly given to turn the boat's head towards Boseham again, and it was not long before they reached its little quay. There the horseman met them, having ridden his horse at full speed, and then Berchthune learnt fuller particulars of the startling news. Cædwalla was only a day's march distant, advancing with a powerful force of West Saxon eorls, and his own veteran band of faithful followers, no longer outlaws, but honoured friends of the king. He was burning to avenge his last defeat and reassert his claim to the throne of the South Saxons.
- 41. This was grave news. Berchthune mounted his horse and rode off at once towards Cissanceaster, directing his followers to come after him as soon as possible. But all this time Ædric and Corman were making the best of their way to Wilfrid. Corman, indeed, when he saw that Ædric was safe, intended going back to look after Father Dicoll, but Wilfrid's men advised him not, and as there was no boat, for they would not lend him theirs, he was compelled to go on. He cast one more lingering, sad look at Boseham, and mourned over his dearly-loved friend, Father Dicoll. But Ædric was delighted; he should not now have to live at Selsea among perfect strangers. After a long ride over a drearily flat country, they came to a clearing amid the gorse and bush; on the other side of this clearing a building, that to Corman and Ædric looked immense, towered aloft over a hamlet of low thatched houses and a few farm buildings. The smell of the sea was all round, and stacks of seaweed filled the air with their peculiar odour. What struck Corman and Ædric, however, was the order and tidiness of everything. The thatched cottages were well thatched, the walls looked well built, and the few people they met all looked better fed and happier than those about Cissanceaster and Boseham. As they got nearer to the large building a solemn sound rose and fell in measured cadence. Ædric had never heard a sound like it, at least not produced by artificial means; it was to him like the wind playing among the tall trees and the sea rolling on the shore mingled with the deep mutter of thunder on the horizon. "What is it, Corman? is it an enchantment?"
- 42. "No, my son; it is the service of vespers in the new church Wilfrid has been building. He has brought over from Rome new wind instruments; and Gregory, the celebrated bishop of Rome, who sent Augustine, the monk, hither, has set new music to the canticles of the church. Thou wilt now be able to see how Christians perform their service of the voice and heart to God." "It is very grand," said Ædric, who had never heard any music more beautiful than the harp, and no singing in combination more than a chorus to some interminable gleeman's tale in verse. They had now got well into the village, and were approaching a long, low, barn-like structure; round the entrance everything was unusually tidy, and some attempt had been made to form a path of shingle and sand, edged with white flints, from the neighbouring beach. In front of this door their guide stopped, Ædric was lifted off the horse-litter by Corman and the other man, and they entered a large room or hall. Ædric had never seen a room like it. The floor was very clean, and a fresh pile of reeds lay near the door, to replace the soiled ones that served as a mat. There was a long table down the middle of the room, and across one end was another table, in the centre of which was a large massive oaken chair; on each side of the table were wooden squares, or trenchers, which served for plates; by the side of these were horn drinking-cups. At the end of the room, opposite the large chair, was a wooden reading-desk, and on this was a splendid manuscript, heavily bound and chained to the desk. Ædric could see that there were some lovely pictures in it, and he longed to examine the volume. He had never seen a book in his life before, and the nearest idea he had ever had of a drawing had been some carvings on a horn which his
- 43. father very highly prized, and some pictured hangings which were treasured among the family's most valued belongings, and which tradition said had been taken in the sack of the haunted ruins at Brædynge. Father Dicoll and the poor monks had no books; they had no parchment, and no paper. Ædric had heard of writing, but it had always been spoken of with awe, for it was considered to savour somewhat of magic. It was therefore with a solemn feeling, as well as one of curiosity, that he looked at the large mysterious volume. At the side of the room opposite the door, and nearly in the middle on that side, was a bright fire. The logs were piled up on iron bars, and a large square of hard trodden clay served as hearth. The smoke from the fire found its way up and out of the hall by an aperture in the roof immediately above it, but, as it did not always take this way out, there was a strong smell of burnt wood and smoke in the room. Ædric and Corman were led up to the bench before the fire, and told that the clerks who were with Wilfrid were at service and they were to wait there until it was over. Ædric felt awestruck at the silence, the neatness, the comfort of everything, but especially at the stillness of the place, the hall of his own home having always been full of noisy domestics, familiar and lazy; the remains of the last, and indeed of several previous feasts, were left on the floor, and the whole place habitually reeked of feasting, rude plenty, and dirt. But here was something very different. Order and cleanliness were visible everywhere. Presently there was a noise of feet outside on the shingle path, and a tall figure entered the room. It was Wilfrid, followed by his two faithful companions, Bernwine and Hildila. Corman at once arose and stood in submissive silence before the great churchman, while
- 44. Ædric tried to get up, but was arrested by the kind voice of Wilfrid bidding him be seated. The boy was at once won by the gentle voice and kind smile of the bishop, but was at the same time much in awe of him. Somehow he seemed so very much farther away from him than Father Dicoll had seemed; it was not that he did not greet him in quite as friendly a way, or with even a kinder smile, but the boy had a feeling that he was a much smaller object, and could not possibly be of any interest to Wilfrid. At the same time there came across him all that Dicoll had said about him, and, with the instinct of a boy who is quick to recognise what is put on or assumed in manner, he felt as if Wilfrid's kindness were a matter of policy, and not a matter of the heart. It is not to be supposed that Ædric could have given these reasons for his awe of him, but in very great awe of Wilfrid he certainly was, and what was even more curious, brother Corman seemed equally in awe of the bishop. As not infrequently happens when very ingenuous, candid natures come in contact with deeper, more intricate, more commanding minds, it seemed to strike both that it was Wilfrid's part to be both kind and sweet in manner, while with Corman himself it was his nature to be so. "My son, thou must be very tired after thy journey," said Wilfrid. "Thy couch is prepared, and supper shall be taken to thee there. I will entrust thee to the care of Father Bernwine, who will make a careful nurse, and see that thou art well cared for. In the morning, if all be well, I will talk with thee. Meanwhile, Good-night, and may the peace of God go with thee." This was all said with such sweet dignity that Ædric, who would much rather have sat up and did not feel at all sleepy, did not
- 45. venture to dispute the arrangement, although at home he would undoubtedly have boisterously done so. He was supported out of the room, therefore, by Bernwine, after taking an affectionate leave of Corman, who remained awaiting the bishop's instructions. CHAPTER XVI. "BLESSED ARE THE PEACEMAKERS." The next day Ædric awoke early. It took him some time to realise where he was. The dim light of morning came in through a narrow aperture in the walls, and he could only just make out surrounding objects. All was very quiet. He could see that he was in a little room, neatly furnished with a wooden settle or stool, and the wooden bed on which he lay. There was a little wooden cross on the wall by the side of his bed, and some writing underneath it, at which Ædric stared, not quite liking it. He thought it must be writing, for it was rather like some marks on the horn at home, and which he had been told were spells. He wished it was rather lighter, or that some one would come, for he could not tell what the runes might do, they might contain some enchantment, it was better not to look at them. Presently he heard the same solemn sound he had heard last night, it sounded very beautiful as it plaintively pealed through the building, now rising in sustained unison, then sinking in deeper notes, appearing to swell and sink and swell again, appealing in mystical utterance to an invisible but all-powerful Being. When the music ceased, Ædric could hear a continuous sound of human
- 46. tongues, then one deep musical voice, followed by a solemn melodious blending of all the tones. Soon after, he heard the noise of steps at his room, over the entrance of which a curtain hung, and in another moment Corman entered, much to the boy's relief. "Oh! how glad I am to see thee, Corman, I feared thou hadst gone away." "I should not have gone without seeing thee first, Ædric; how hast thou slept?" "Very well. I feel much better, I believe I shall be able to walk without any help, to-morrow, if I am allowed to practise a little to- day—but what was that sound? what have they been doing? I never heard anything like it." "That was the morning service, or matins, and I have just come from it. But thou hadst better get up now, and I will help thee into the Refectory, where we are all going to have breakfast." When Corman and Ædric entered the large room or hall, into which they had first come the evening before, they found the room nearly full. Wilfrid was at the head of the table, on each side of him were Bernwine and Hildila, while all down the long table were a few monks, some lay domestics, and several boys, who all looked curiously at Ædric. One of the monks led Corman and Ædric to their vacant places, and then grace was said by another monk at the lower end of the table, after which all sat down, and the same monk who had said grace, began to read out of the beautiful book that had so attracted Ædric's attention the night before. The breakfast consisted of a portion of fish to each person, and a portion of oatmeal porridge made with water. There was water to
- 47. drink, but at Wilfrid's table there was a jug of milk, of which, however, the Bishop only took very sparingly, but he sent it down to Ædric, and another monk who seemed delicate, bidding them take it for their bodily comfort. The fish had been caught by Wilfrid himself, who had taught the ignorant South Saxons how to supply themselves with this wholesome food, and, like many men remarkable for their intellectual gifts, he was especially pleased with the success of his skill in the gentle craft. No word was spoken during the meal, all listened attentively to the reading of the monk. He was reading from "The Dialogues" of Pope Gregory the Great, but Ædric naturally did not understand a word, as it was all in Latin; when Corman afterwards told him the marvellous tales that the monk had read, he wished much that he could have understood it, and longed more than ever to look at the pictures, and made up his mind he would like to learn to read. When all had finished, the reader closed the book and said grace, after which he sat down and had his own breakfast, while the rest dispersed. It appeared that each man had his allotted task; some went to the outhouses whither the platters and other appliances of the breakfast table were taken, and were there washed up; one of the lay brothers winding up a bucket of water from the well hard by, and heating it in a copper. Others went to a tool-house, and taking their hoes and mattocks went out to the garden on the south side of the little settlement. The choir boys were taken off to the church and were there taught general knowledge, as well as music, by Bernwine. Hildila took two or three monks with him and they carefully practised writing under his instruction.
- 48. Wilfrid beckoned to Corman to bring Ædric up to him. The boy felt very shy when he saw the clear piercing grey eyes of the celebrated Bishop searching him through and through. For in Wilfrid's face there was that presence of a will, which is always so marked in men who have been great in the world, and this will makes its presence felt without a word being spoken, as the needle, when magnetised, is powerless to resist the attraction of the mysterious pole. "My son, brother Corman has told me all about thee. He tells me how patient thou hast been under suffering, and how thou hast been brought to wish to lead a better life. Thank God for thy pain, for by it thou hast been enabled to learn the way of salvation, and mayst be intended for a blessed purpose, even the awakening of thy people from the dark night of Paganism to the glorious light of the Gospel." Ædric looked timidly at Wilfrid: he did not know what to say, he could not talk to him as he had done to Father Dicoll and brother Corman. He felt he could only learn by hearing, not by questioning, which, to a boy, is so much the preferable way, but which, unless carefully directed, leads many times to a desultory and fruitless end. Wilfrid went on, seeing that the boy was listening: "Thou wilt be able to learn many things here. When thou art thoroughly taught in all that is necessary, thou shalt be baptised; and when thou hast quite recovered, thou canst return to thine own land and teach thine own people. For what more beautiful or holy object canst thou have in life than the hope of meeting those who have been brought to eternal life by thy means? Think what a blessed thing it would be if the Almighty should employ thee as His messenger. And be not