Decide for Dummies

DECIDE for Dummies
Juncal Alonso, Leire Orue-Echevarria (TECNALIA)

DECIDE for Dummies
 This presentation is a guide to the usage of the different components
developed in the context of DECIDE H2020 Project.
 A video of the demo is also available in the DECIDE H2020 webpage
here.
 The presentation is organized as follows:
 Chapter 1: DECIDE DevOps framework access
 Chapter 2: The used example: Sockshop application
 Chapter 3: DECIDE DevOps framework usage
• Chapter 3.1: DECIDE extended DevOps framework
• Chapter 3.2: Architectural design
• Chapter 3.3: Deployment optimization
• Chapter 3.4: Multi-cloud SLA creation
• Chapter 3.5: Resources contracting
• Chapter 3.6: Components deployment
• Chapter 3.7:Components and underlying resources monitoring
• Chapter 3.8: Multi-cloud application adaptation
• Chapter 3.9: Billing
GA 731533 (c) DECIDE Consortium 2

DECIDE DevOps framework access
DECIDE integrated framework access and installation guidelines

DevOps framework access
 The DevOps framework is the integrated
framework for all the DECIDE tools.
 To follow this DECIDE for Dummies tutorial the
DevOps framework is accesible here:
 http://85.91.40.245:8084/
 It can also be installed on premise. For that
please follow the corresponding guidelines on
the annex: D2.8-Final Integrated DevOps
framework.

The used example: The Sockshop
application
Introduction to the used sample multi-cloud based application

The Sockshop application
 During this tutorial the examples shown will be based on the Sockshop
application.
 The SockShop App1, is a loosely coupled microservices demo application. It
simulates the user-facing part of an e-commerce website that sells socks. It is
available as open source software and has been developed with the intention to
aid in demonstrating and testing microservices and cloud native technologies.
 The Sock Shop app is designed to provide as many microservices as possible. The
microservices are defined by functionality required in an e-commerce site and
are loosely coupled.
 Sock Shop microservices are designed to have minimal expectations, using DNS
to find other services. The Application uses a message broker for sending
messages using queues.
 All services communicate using REST over HTTP. Furthermore, the Sock Shop app
is polyglot being built using Spring Boot4, Go kit5 and Node.js6 and is packaged
in Docker containers.
 During this tutorial the example’s will be based on the Sockshop application with
10 microservices
1 https://microservices-demo.github.io/

DECIDE extended DevOps framework
Integrated framework to support the development and operation of multi-cloud applications

DevOps framework: theory
Development and Operations
Agility Stability
DevOps

 There is not a common and agreed definition for DevOps
 DevOps is a set of practices intended to reduce the time between
committing a change to a system and the change being placed into
normal production, while ensuring high quality. [Bass, Len; Weber,
Ingo; Zhu, Liming. DevOps: A Software Architect's Perspective.
ISBN 978-0134049847.]
 DevOps represents a change in IT culture, focusing on rapid IT service
delivery through the adoption of agile, lean practices in the context of
a system-oriented approach. DevOps emphasizes people (and
culture), and seeks to improve collaboration between operations and
development teams. DevOps implementations utilize technology —
especially automation tools that can leverage an increasingly
programmable and dynamic infrastructure from a life cycle
perspective. [gartner, https://www.gartner.com/it-glossary/devops]
DevOps

 All definitions present commom characteristics
DevOps seek to bring
closer the interaction
between Dev and
Ops, using similar
tools
Continuous delivery pipeline
• Continuous integration (CI)
• Continuous Quality (CQ)
• Continuous Delivery(CD)
DevOps

DevOps
 DevOps is represented as an infinte loop

https://blog.rackspace.com/uk/rackspace-finds-clear-business-benefits-increasing-use-devops-rapid-software-
development-deployment
Efficiency
Customer
satisfaction
Uptime of
the app
Agility More value:
new
capabilities
Productivity and
satisfaction of
the team
Conversion of
new leads
DevOps: benefits

DevOps: barriers
Source: www.quali.com
Culture Fail to
automate a
part of the
process
Legacy
applications /
infrastructure
Application
complexity
No clear
plan of
DevOps
Tools Management of
environments
Training Commitment
of Mgmt
Budget

DevOps framework: multi-cloud
 Multi-cloud applications with strict NFR, more
specifically, performance, availability, location
and cost concerns
+ + +
+ + =
Multi-cloud in most literature Multi-cloud for us
Multi-cloud

DevOps framework: multi-cloud
Same CSPs, Different location
Different CSPs
Different resources / services in the
same CSP and location
Multi-cloud

DECIDE Approach
Non-functional
requirements
gathering and
prioritization
Architectural
Design
Development,
integration and
testing
CSPs and
microservices
profiling and
classification
Optimization (+
CSP Discovery)
Application
MCSLA
definition
Deployment
CSPs service
contracting
Deployment
configuration
provisioning
Application
MCSLA and NFR
Monitoring
CSPs SLA
monitoring
Self-adaptation
and automatic
(re-)deployment
Support in the SDLC and SOLC in an integrated manner
ARCHITECT OPTIMUS ACSmI ADAPT DevOps FW

DevOps Framework: in practice
 Register yourself in the DevOps framework. Click
Register

DevOps Framework
 Complete the
details in the
next screen
and click
Register

DevOps Framework
 Login using
the email and
the password
selected.
 When you
login, you will
see the
following
screen

DevOps Framework
 You can edit your profile at all times by clicking

DevOps Framework
 The following screen appears when clicking
 You can change here your profile and include the credentials
needed for the Docker registry. This is necessary to run ADAPT
DO

DevOps Framework
 In the Credentials Tab insert the information related to the
Docker registries where you have your containers
 Upload also the certificate that your Docker registry needs

DevOps Framework
 You can create a new application by clicking on
the button new application

DevOps Framework
 And you will get this screen

DevOps framework
 Complete the following fields:
 Name: insert the name of the application
 Description: insert a description of the application
 Git repository: include the URL of the git repository
where the DECIDE.json will be stored (see next for
more explanations on the DECIDE.json)
 Git token: the token needed to access the git
 High technological risk: if this option is selected, the
automatic redeployment will not be executed
Import Description from DECIDE Project: in the case
you have already a DECIDE.json from a previous
project, you can import it here.
 When you are done, click

DevOps framework
 Clicking on you can add the microservices.
 The following data must be inserted:
Name: name of the microservice
Programing language: the language in which it was
developed
Stateful: if the microservice has state or not
Public IP: if the microservice should have a public IP
(e.g. in the case it is a user interface)
Tags: this is used for a later stage, for the NFRs and the
monitoring. By default, the tag ‘application’ is included,
but you can also insert the tag that indicates one of the
microservices.
 When you are done, click

DevOps framework
 Now include the NFRs.
 The NFRs availability,
performance location,
and cost are later on
monitored and used in
ARCHITECT for the
recommendation of
the patterns
 The NFR scalability and
legal level are only
used by ARCHITECT for
the recommendation
of the patterns
NFR

• Availability level: high,
medium or low.
Depending on the level
selected the number of
suggested patterns by
ARCHITECT will vary
• Availability (%): that is the
percentage of availability
for the whole application
(by default) and in the
case additional tags were
inserted when creating
the microservice, these
values will also applied to
them. This value will be
used later on for
monitoring purposes
• Tags: select here if the
NFR is defined at
application level or at
microservice level
DevOps framework
NFR

DevOps framework
 When clicking the details of the
selected NFR will appear below
NFR

• Performance level: high,
medium or low.
Depending on the level
selected the number of
suggested patterns by
ARCHITECT will vary
• Response time (ms): the
maximum time allowed
for the whole application
(by default) to respond. In
the case additional tags
were inserted when
creating the microservice,
these values will also
applied to them. This
value will be used later on
for monitoring purposes
NFR is defined at
microservice level
DevOps framework
NFR

• Scalability level: high,
medium or low. Depending
on the level selected the
number of suggested
patterns by ARCHITECT will
vary
• Scalability value
(request/sec): the
maximum numbers of
requests per second
allowed for the whole
application (by default. In
the case additional tags
were inserted when
creating the microservice,
these values will also
applied to them. This value
will not be used later on for
monitoring purposes
• Tags: select here if the NFR
is defined at application
level or at microservice
level
DevOps framework
NFR

• Location type: single
location, single country,
cross-border. This will be
used later on for the
selection of the cloud
offerings.
• Region: Europe, Asia, north
America, south America.
The region in which the
cloud service offering
should be located. This
value will be used later on
for monitoring purposes
• Zone: the zone where the
cloud service offering will
be located (e.g. Germany,
France, US …). This value
will be used later on for
monitoring purposes
• Tags: select here if the NFR
is defined at application
level or at microservice
level
DevOps framework
NFR

• Cost level: high, medium
or low. Depending on the
level selected the number
of suggested patterns by
ARCHITECT will vary
• Cost: The maximum cost
you would be willing to
pay. This value will be
used later on for
monitoring purposes
• Currency: Euro, pound,
dollar
NFR is defined at
microservice level
DevOps framework
NFR
Euro, pound, dollar

• Legal level: tier 1, tier 2
and tier 3. This value will
be used for the selection
of the cloud offerings
DevOps framework
NFR
The explanation of the legal levels are found next

DevOps framework
 Legal level tier 3: basic legal safeguards
Basic general safeguards (company registration,
presence of representative if relevant, presence of a
DPO or data protection officer, DPA and transfer
mechanism where relevant)
Basic contractual safeguards (liability, confidentiality,
exit options, ADR)
Basic GDPR compliance (basic contractual guarantees
are present but may be faulty depending on
interpretation, conditions may not be ideal)
No certification
No adherence to approved self-regulatory instruments

DevOps framework
 Legal level tier 2: substantial legal safeguards
Basic general safeguards (company registration, presence of
representative if relevant, presence of a DPO or data
protection officer, DPA and transfer mechanism where
relevant)
Substantial contractual guarantees for the most relevant
aspects (liability, confidentiality, exit options, ADR), basic
guarantees for others
Substantial GDPR compliance (balanced and substantial
contractual guarantees, issues or challenges are unlikely,
although conditions may not be ideal)
Valid basic certification (ISO 27001 or equivalent) covering
the service
No adherence to approved self-regulatory instruments

DevOps framework
 Legal level tier 1: strong legal safeguards
 Basic general safeguards (company registration, presence of
representative if relevant, presence of a DPO or data protection
officer, DPA and transfer mechanism where relevant)
 Strong contractual guarantees for the most relevant aspects
(liability, confidentiality, exit options, ADR), substantial
guarantees for others
 Strong GDPR compliance (strong contractual guarantees with
favorable conditions of the Cloud user/controller)
 Valid basic certification (ISO 27001 or equivalent)
 Valid Advanced cloud-specific certification that meets all of the
27 security objectives of the Cloud Certification Schemes
Metaframework as defined by ENISA covering the service
 Adherence to approved self-regulatory instruments

DevOps framework
 The final set of selected NFRs are as follows.
Then click ‘Save’

DevOps framework
 The DECIDE.json is created and stored in the git

Application description
40GA 731533 (c) DECIDE Consortium
 Each tool shares needed information by
pushing and pulling from a dedicated
repository.
 Information model specific to the DECIDE
Application that is stored in a repository (e.g.
Git) and is accessed by each tool. It is a JSON
file
 Interoperable mechanism

Application description
41GA 731533 (c) DECIDE Consortium
 It looks like this

DevOps framework
 Now the dashboard has changed and it shows
the main characteristics of our application

DevOps framework
 In the dashboard you can see:
 the NFRs selected and their values
 the microservices of the application and the tags
 The integration with Jenkins and SonarQube
… More fields

DevOps framework
 DECIDE allows to integrate with CI tools such as
Jenkins

DevOps framework
 Now that the
basic data of
the application
has been
included, the
left menu
shows in green,
the actions that
can be carried
out

DevOps framework
 In this case, the next step is the architectural
design but also the cloud services Discovery
could be used.
 Cloud services can be discovered at any point
in the workflow.

Architectural design
Architectural design for multi-cloud applications with relevant NFRs using ARCHITECT

ARCHITECT: theory
 Based on the NFRs and the values inserted,
DECIDE ARCHITECT recommends a set of
patterns
 ARCHITECT is distributed in two flavours:
Eclipse Plug inWeb application
+
Integrated with OPTIMUS
Integrated with DevOps
framework

ARCHITECT: theory
Pattern Classification
Fundamental (4)
Development (33)
Deployment (20)
Optimization (12)
Pattern Description
Name / Type / Context
Problem / Solution
Model
NFR Impact
Related Patterns
DECIDE ARCHITECT
 There are patterns defined for each phase of
the application plus a set of fundamental
ones.
 All patterns are described the same way

 The patterns are recommended based on the
NFRs values
ARCHITECT: theory
DECIDE ARCHITECT: Pattern recommendation
Pattern
recommendation
Software
architecture
Deployment
simulation
Development
patterns
Optimization
patterns
Deployment
patterns
Deployment
requirements
NFRs

ARCHITECT: theory
DECIDE ARCHITECT: Pattern recommendation
 The inference and recommendation
mechanisms are shown below:

ARCHITECT: in practice
Recommended patterns
All patterns avaiable

 The recommended patterns can be grouped
by:
 alphabetical order
 non-functional requirement (general, availability,
location, performance,…)
 category (fundamental, development,
deployment, optimization)

 Clicking on each pattern the following
information is shown:

 Select the patterns that you consider relevant and
‘Confirm Selection’
 The application description file is updated with the
selected patterns (field “selected” is true)

 ARCHITECT is also offered as an Eclipse plugin. The way it
Works is the same as the one integrated in the DevOps
framework
DECIDE ARCHITECT: in practice

Deployment optimization
Deployment configuration optimization for specific NFRs and cloud resources classification using
OPTIMUS

DECIDE OPTIMUS: theory
 OPTIMUS has the following functionalities:
 Classifies the components of the multi-cloud
application
 Matches these components with existing cloud
service offerings
Proposes deployment scenarios

OPTIMUS: theory
Eclipse Plug inWeb application
+
Integrated with
ARCHITECT & OPTIMUS
Simulation
REST service
DevOps
Framework
(redeployment)
Integrated with the
DevOps framework
 OPTIMUS is distributed in two flavours:

OPTIMUS Theory
 Classification of the microservices based on the
requirements for each component :
Computing Needs for
public access
StorageDatabase
Main microservices Detachable Resources

OPTIMUS Theory
 Classification of the cloud offerings in services
classes
 Each service class has specific attributes to
characterize it.
Virtual Machines DataBaseStorage

OPTIMUS in practice
 In the tab ‘deployment simulation’ the
functionalities provided by OPTIMUS are
shown.
 Microservice Edit
• allows to add a detachable resource to the microservice
• Allows to add more information about the requirements
needed to deploy that microservice on a cloud service
offering
Simulate
• Provides the user with the different deployment
topologies

OPTIMUS in practice
 Edit the
microservices
by selecting
each
microservice
and including
the
infrastructure
requirements

OPTIMUS in practice
 You can add
also
detachable
resources
(databases)

OPTIMUS in practice
 When going to the tab ‘Simulate’ the following screen is shown
 A preferred cloud provider can be selected. If done so, then
only cloud service offerings from that CSP will be selected. To
do this, OPTIMUS calls ACSmI Discovery
 When you click ‘Simulate’ the algorithm starts (see next)

OPTIMUS theory
DECIDE OPTIMUS: Algorithm
Simulation
REST service
Uses NSGA-II Genetic Algorithm for generating optimized deployment schemas (*).
The steps are:
Initialize
Initial population based on the “OPTIMUS
problem” definition. Population is a set of
solutions (deployment schemas)
Evaluate Level of fulfillment of the FRs and NFRs
established
Select
Comparisons between new Solutions and existing
ones determine the solution to be kept
Crossover &
mutation
Creation of new Solutions improving the existing
ones
Stop when the number of final solutions is less tan five
(*) MOEA Framework, a free and open source Java library for developing and experimenting with
multiobjective evolutionary algorithms

OPTIMUS in practice
 The result is shown next. Select the
configuration you wish
Select and save the
schema

OPTIMUS in practice
 A new section in the DECIDE.json has been
generated. It’s under ‘schema’

DECIDE OPTIMUS
Eclipse plugin
Eclipse Plug in
Integrated with OPTIMUS
Clone Project
Plugin editor/classification
Plugin editor/simulation
Plugin editor/schemas
 The eclipse plug in has the same functionality

Multi-cloud SLA creation
Support for Multi-cloud SLA creation

MCSLA
REST serviceWeb application
+
Stand alone
+
DECIDE DevOps Dashboard
ISO/IEC 19086

MCSLA
MCSLA
 Requests information from ACSmI for all CSLAs
 It also allows to:
Edit MCSLA View MCSLA
 Definition of SLOs and SQOs
 Aggregation of SLOs (*)
(*) Performance and Availability

MCSLA
 The MCSLA shown is the aggregation of the SLAs
from the selected CSPs

MCSLA
 The values suggested are the result from the
aggregation but these can also be changed if wished

MCSLA
 To save the
changes:
click
commit. The
application
description
has been
updated in
the section
mcsla

Resources contracting
Cloud resources contracting using ACSmI contracting

ACSmI contracting overview
 ACSmI contracting offers the functionalities to contract
the resources which are selected by OPTIMUS and
noted in the DECIDE.json (schema section).
 ACSmI contracting programatically connects to the
available CSPs contracting functionalities and performs
the different contracts automatically (without human
intervention).
 ACSmI contracting reads the schema in the DECIDE.json
and automatically contracts the resources. These
resources need to be:
Registered in the ACSmI registry (transparent to the user)
Available in the ACSmI contracting catalogue (transparent to
the user)

ACSmI contracting overview
 The user needs to have an account in ACSmI
contracting to contract the resources.
 The user can have an Amazon personal account
and provide it to ACSmI contracting in case he/she
wants to use this account.
 Once the contracts are stablished, ACSmI
contracting includes the information about the
resources contracted in the DCIDE.json under the
“virtual machines” section
 With this information ADAPT DO can provision the
contracted resources.

ACSmI contracting in practice
 Once the MCSLA is created the DevOps team member can
contract the resources to be used for the deployment. These
resources are already selected and the ACSmI contracting tool
guides the user through the contracting process:

 The user needs to have an account. Only one account can be related to one
DECIDE application. Cloud service contracting register form.

 Cloud service contracting welcome page if the
user is already logged

 All the proposed resources and their characteristics are shown (in this case
Amazon). If the user has an Amazon account, it can be used for the current
contracting, otherwise ACSmI contracting offers the possibility to do it
through the ACSmI platform directly, who has a specific framework contract
with Amazon cloud provider.

 Characteristics of the proposed contract

 Once the contracts are stablished the user gets a success
response and the contracts can be checked under the contracts
tab.

 Current active contracts

Components deployment
Automatic Multi-cloud components deployment (microservices) in contracted cloud resources
Using ADAPT Deployment Orchestrator (ADAPT-DO)

ADAPT DO overview
 ADAPT DO provisions the machines contracted, install
all the needed software to deploy the Docker images,
downloads all the images from a Docker registry and
deploys them in the corresponding cloud resource (see
flow description in the following slides).
 ADAPT DO sets up all the communications between the
containers (with the information from the developer).
 ADAPT DO reads the schema and the virtuals machines
section in the DECIDE.json to automatically deploy the
components. The user needs to provide information
about:
The components communication (ports, etc)
The components source (images name, Docker registry
information)

ADAPT DO overview
 ADAPT DO installs the agents needed for the
monitoring of the resources and monitors the
application through ADAPT monitoring and the
resources through ACSmI monitoring.

ADAPT DO flow
ADAPT
Deployment Orchestrator
DevOps
Framework /
State Machine
Git repo
Applicatio
n
descriptor
ADAPT parses the Application Descriptor
and detects how many VMs must be started and on which CSP

ADAPT DO flow
ADAPT
DevOps
Framework /
State Machine
VM
VM
VM
VM
VM
VM
ACSMI
Git repo
Applicatio
n
descriptor
ADAPT invokes the ACSmI broker to ask for the provisioning of VMs
on the target CSPs
ACSmI applies the provisioning requests to the proper CSPs

ADAPT DO flow
ADAPT
DevOps
Framework /
State Machine
VM
VM
VM
VM
VM
VM
ACSMI
runtim
e
security
runtim
e
security
runtim
e
security
runtim
e
security
runtim
e
security
runtim
e
security
Git repo
Applicatio
n
descriptor
When VMs are ready (have a public address) ADAPT accesses them
and installs on each one the needed runtimes and configurations

ADAPT DO flow
ADAPT
OPTIMUS
VM
VM
VM
VM
VM
VM
ACSMI
s
runtim
e
security
runtim
e
security runtim
e
security
s
s
runtim
e
security
runtim
e
security
runtim
e
security
s
s
s
Git repo
Applicatio
n
descriptor
Microservice CSP VM
microservice1 CSP1 VM1
Finally ADAPT can deploy the microservices according to the
Application Description deployment schema

ADAPT DO in practice
 STEP 1- ADAPT DO information: This
information is automatically loaded from the
DevOps framework and the DECIDE.json. The
user only needs to continue to the next step.
Click to continue to the
next step

 STEP 2 – Resources information: This information is
automatically loaded from the DECIDE.json. If
necessary, the user can change it (if the resources
have been contracted manually).
Click to continue to the
next step

 STEP 3 – Containers information: This information
needs to be completed by the developer/operator of
the application (in red mandatory fields). This
information includes:
Container name: To be decided by the developer
Image name: As it is named in the registry loaded in the
profile1.
VM_Name: It is loaded from the DECIDE.json
Tag: Image tag to be downloaded from the registry
Docker registry: Select from the list the registry where the
Docker images will be downloaded by ADAPT. This registry is
configured in the DevOps user profile information1.
1See DevOps FW in practice profile creation

 Port mapping: This information needs to be provided by the
developer/operator of the application. It corresponds to the
actual mapping of the host (resource) ports and the containers
ports.
 Host mapping: This information needs to be provided by the
developer/operator. It corresponds to the add-host command in
Docker CLI, that configures an specific service in a container to
reach external hosts.
 Volume Mapping: This information needs to be provided by the
developer/operator. It corresponds to the volumes to be added
to the container for persistent storage of data.
 Safe methods: This information needs to be provided by the
developer/operator. It corresponds to the REST method against
which ADAPT Monitoring will monitor the availability of the
component (application level). This method needs to be safe (it
should not have side effects in the application performance).

 Reverse Proxy Endpoints: This is a mechanisms which allows to set up a
reverse proxy, so that microservices reference each other via name,
while the actual adresses are stored in a remote location and always
kept up to date during re-deployments. When a microservice refers to a
another microservice, it requests the proxy for it by name and the
proxy knows the target service that is responding at a specific moment.
 Command parameters: These are parameters to be passed to a
container for the EntryPoint command (see EntryPoint Docker
command).
 Entrypoint: This is the command to be run in the EntryPoint command
(see EntryPoint Docker command).
 Networks: This is the value for the network option when starting the
container with Docker run.
 Environment: This is the definition of environment variables when
starting a container with Docker run.
 Add consul service/Consul service port

 Generic container info

 Port mapping and host mapping information

 Volume mapping and safe methods.

 Reverse proxy, command paramenters,
Entrypoint, Networks and Environment.

 Consul service port

 STEP 4 – This step submits the preparation
information to ADAPT DO and launches the actual
deployment process including (for more information
see ADAPT DO Flow):
Infrastructure initialization, planning and application
(following the information of the Application description).
Services initialization, planning and application (following
the information of the Application description).
 The user can perform the different steps one by one or all in
one shot.

 STEP 4 – One by one option: Each of the phases are
manually launched by the user. When one by one option is
selected, first the preparation
step needs to be submitted
1
2 3

 STEP 4 – One shot option:The whole process is
launched at once.When one shot deployment, the whole process is launched
with the corresponding option. The next phases are
automatically launched when the previous one is finished
1

 STEP 4 – In both cases the “submit preparation step”
or “one shot deployment (all in one deployment
step)” will init the process by accepting the proposed
data: The user needs to accept the
proposed data submission

 STEP 4 – Once the actual deployment process has
started the status of the different resources and
microservices can be checked by the green ticks and
the circles in the UI.

 Once the deployment process is finished the user can check the deployed
application (in this case the sockshop) in the corresponding virtual machine (public
IP in ADAPT UI or dockerHostPublicIp in the virtual machines section in the
application description), accesing the desired component. In the Sockshop case , we
need to access the front-end container.
The front-end is deployed in the
VM1_Amazon_Ubuntu16_04

 Once the deployment process is finished the user can check the deployed
application (in this case the sockshop) in the corresponding virtual machine (public
IP in ADAPT UI or dockerHostPublicIp in the virtual machines section in the
application description), accessing the desired component. In the Sockshop case, we
need to access the front-end container.

Components and underlying
resources monitoring
Proactive Multi-cloud components (microservices) and underlying resources (virtual machines)
monitoring and continous SLA assessment

Generic monitoring overview in
DECIDE
 Monitoring approach in DECIDE
ADAPT
monitoring
ACSmI
monitoring
Out of
scope
(*) Source: Production Ready Microservices O’Reilly
Layer 4: Microservices
Layer 3: Application platform
Layer 2: Communication
Layer 1: Virtual Hardware
Out of
scope

Generic monitoring overview in
DECIDE
 Monitoring stack in DECIDE
Composed metrics
visualization
Composed metrics
generation
Raw metrics storage (TSDB)
Agents for metrics collection
ACSmI
monitoring
ADAPT
monitoring
ADAPT
monitoring
ACSmI
monitoring
ACSmI
monitoring
ADAPT
monitoring
Composed metrics
assessment
ACSmI
monitoring
ADAPT
monitoring
ADAPT
monitoring
MCSLA
editor

ADAPT Mon overview
 ADAPT Monitoring (ADAPT Mon) monitors the different components
of the multi-cloud application through the safe methods introduced
by the developer (see ADAPT DO) against the NFRs established in the
initial phase (see DevOps framework): Availability and/or
performance.
 ADAPT Mon also monitors the fulfillment of the established NFRs
(availability and/or performance) at application level if any (see
DevOps framework).
 NFR monitored metrics are generic and based on the ISO19086
standard.
 Availability = MTBF/(MTBF+MTTR)
 Performance = Response time
 When any of the stablished NFRs is violated, ADAPT Mon triggers the
Violation Handler which is the component in charge of managing
violations.
 Apart from the application layer monitoring, ADAPT Mon also
launches the infrastructure monitoring, calling ACSmI mon.

ADAPT Mon in practice
 Once the multi-cloud application is deployed the
monitoring tabs are automatically activated, with the
corresponding selected metrics (performance).

 Once the multi-cloud application is deployed the
monitoring tabs are automatically activated, with the
corresponding selected metrics (availability).

 When any of the established NFRs is violated, ADAPT Mon
triggers the Violation Handler which is the component in
charge of managing violations.
-> Alarms raised Visual feedback

ACSmI Mon overview
 ACSmI Monitoring (ACSmI Mon) monitors the resources where the
different components are being deployed.
 ACSmI Mon monitors the compliance at run time of the following NFRs
(stated in the CSPs SLAs). monitored metrics are generic and compliant
with ISO19086
 Location = Actual location of the VM
 Availability = MTBF/(MTBF+MTTR)
 Performance = Disk, Memory and CPU usage
 Cost = Actual cost and usage records of each resource.
 When any of the established SLOs is violated, ACSmi Mon triggers the
Violation Handler which is the component in charge of managing
violations.
 When the violation is detected by ACSmI mon the ACSmI registry is
updated so that the user (and other DECIDE tools like Optimus) are
aware of this violation and can perform the required actions when
looking for services in the next deployment.

ACSmI Mon in practice
 ACSmI Mon does not include a graphical interface for metrics
monitoring but the violations can be checked in the ACSmI
Discovery user interface.
Violations detail
Violations occurred

Multi-cloud application adaptation
Semi-automatic Multi-cloud application adaptation if a violation occurs

Violation Handler overview
 The Violation Handler (VH) is the component which
manages the violations.
 It receives the violations from ADAPT mon or ACSmI mon
and performs the following steps:
 Notifies the developer so he or she is aware of the violation.
 If the technology risk of the application is low (see DevOps
framework description) the VH triggers a new deployment
process which includes the following automatic tasks:
• Call OPTIMUS for a new deployment schema
• Create a new MCSLA
• Contract the new resources selected by OPTIMUS
• Deploy the components in the new resources and stop the old ones.
• Set up the monitoring components for the new deployed
components/resources.

Violation handler in practice
 The developer receives the email from the VH and the re-
deployment process is launched (if the technological risk is
low).
The re-deployment process is
triggered with the call to OPTIMUS
who writes the new schema in the
application description

Violation handler in practice
 The redeployment process is launched automatically if the
technological risk is low. The different DECIDE components are
subsequently automatically executed until the new
redeployment takes place. Each component updates the
required information in the application description.
OPTIMUS writes the new schema
OPTIMUS updates the history
New MCSLA is created
New contracts are established
The application is deployed in the
new resources

Billing
Proactive Multi-cloud components (microservices) and underlying resources (virtual machines)
monitoring and continous SLA assessment

ACSmI Billing overview
 ACSmI Billing keeps track of the usage of each cloud resource
and the corresponding costs.
 The user can access the ACSmI billing to check the active
invoices and the usage records.

Decide for Dummies

More Related Content

What's hot (19)

Similar to Decide for Dummies (20)

Recently uploaded (20)

Decide for Dummies

Editor's Notes