Making Sense Of Cloud Computing - by Mark Rivington

Making Sense of Cloud Service Computing Mark Rivington, VP Technology, Nimsoft December 2009 Keys to Effective Cloud Service Management

Agenda What is Cloud Computing? Business Drivers for the Cloud The “5-4-3” of Cloud Computing 5 Cloud Characteristics 4 Deployment Models 3 Service Offering Models Managing services in the Cloud Balance of responsibility Specific monitoring activities Cautionary considerations

What businesses are saying... Source: http://geekandpoke.typepad.com

What is Cloud Computing? Not to be confused with … Grid computing, Utility computing, Autonomic Computing, Service Orientated Architecture, Virtualization, Web Services and all previous computing models Definitions are forever changing In March 2007, Dell applied to trademark the term "cloud computing" ( U.S. Trademark 77,139,082 ) in the United States . The "Notice of Allowance" it received in July 2008 was cancelled on August 6 , resulting in a formal rejection of the trademark application less than a week later. Cloud computing is Internet ("cloud") based development and use of computer technology. It is a paradigm of computing where dynamically scalable and often virtualised resources are provided as a service over the Internet. Users need not have knowledge of, expertise in, or control over the technology infrastructure "in the cloud" that supports them. -- Wikipedia.org

Business Drivers for the Cloud Relative to classical data center centric IT Escalating cost of data center ownership Asymmetric workloads, unknown workloads Surge computing Stepwise rather than linear capacity increases Difficult to manage and plan Increasing Risk associated with data center loss Business realignment to core Not in the business of managing data centers Maturity and value of cloud offerings constantly increases

3 Service Offering Models - Infrastructure as a Service (IaaS) - Platform as a Service (PaaS) - Software as a Service (SaaS) The “5-4-3” of Cloud Computing 4 Deployment Models - Private Cloud - Public Cloud - Hybrid Cloud - Community Cloud 5 Cloud Characteristics Elastic resources IT service centric approach Consumption based metering Self-service ubiquitous network access Location independent resource pooling

Five Characteristics of Cloud Computing Elastic resources Resources can quickly grow & shrink, appearing to be infinite IT service centric approach Able to run business services or applications for a specific, timely purpose Source: http://blog.washingtonpost.com

Five Characteristics of Cloud Computing Consumption-based metering – “Pay-as-you-go” Model Service usage is measured No upfront or long-term commitments Source: http://thelittlechimpsociety.com

Five Characteristics of Cloud Computing Self-service ubiquitous network access Accessible from anywhere on any device on-demand Allows users the ability to upload, build, deploy, schedule, manage & report on their business services on demand Location independent resource pooling Shared resources assigned according to multitenant demands Resources are virtualized & can change dynamically Source: http://ryan2point0.wordpress.com

Four Cloud Deployment Models Private Cloud Similar to co-location but with elastic resource Public Cloud Cloud Computing popular delivery model - truly Public Hybrid Cloud Combination of Public and Private with overflow capabilities Community Cloud Community access, such as gmail, yahoo, salesforce...

Private Clouds Massively virtualized infrastructure in private datacenters Promoted and provided by key virtualization vendors – VMware, Citrix, Microsoft, IBM, SUN, Eucalyptus (open-source) As well as the storage vendors – EMC, Hitachi, Netapp and others Likely to be very strong wave in the coming years (2010  Test, 2011  Production) Hybrid private-public cloud deployments “ VMware’s focus is on enabling our customers to run their datacenters as internal clouds and operate in a far more flexible and cost-efficient way,” said Paul Maritz, president and chief executive officer, VMware . “Our customers want the plumbing to disappear – in the datacenter, on the desktop and in the cloud – so they can focus their staff time and IT budget on delivering business value. They want cloud-like services so they can act as hosting providers to their internal customers. Our Virtual Datacenter Operating System Initiative will accelerate customers down the virtualization path so that they can run their IT as an internal cloud service. The VMware vSphere generation of products, which are currently in development, will be a new class of software that delivers on this strategy. And, as customers become cloud-enabled, they will have the flexibility to securely and efficiently expand their internal clouds to tap the resources offered by external service providers through our VMware vCloud Initiative. I am excited to share the progress we’ve made on our initiatives at VMworld Europe.” VMWare press release – February 2009

Monitoring Private Clouds “ Business as usual” for “good” Monitoring Solutions Broad technology coverage Extremely Scalable Zero touch provisioning of monitoring Highly dynamic monitoring qualities Now you see it, now you don’t! Virtualization platform performance data collection VMWare, Microsoft, Citrix XEN, IBM, SUN and O/source Standard end-user performance monitoring Business Service Management Relating highly dynamic infrastructure to business services

Service Offering Model (1 of 3) – SaaS Characteristics Specific services provided as hosted solutions Driven by standard service feature/function requirements Consumers use the service via an application interface & configure settings Very rapid start up times Usage is normally subscription based on key application parameters e.g. data storage size, # of users, etc. Compelling alternative to internal IT particularly with a lack of resources & capability

Management Responsibility – SaaS Degree of Management Responsibility Consumers Service Providers Capacity Change Configuration Data protection Compliance Event Problem Security Asset Server OS Application Service Database Network Examples: Google Apps PayPal Payment Services SalesForce.com

Monitoring SaaS Clouds End User Monitoring Synthetic Transaction Monitoring Custom support with “record-n-playback” Real User Monitoring Inserting an appliance on network edge to passively measure outgoing requests & responses URL and Web Service Response monitoring Scalable and highly effective Application specific instrumentation Application dashboards URL data gathering Metrics gathered through available APIs Vendor specific Service Level Agreement reporting Automatically generated SLA reports Business Service Management reporting Compliance to SLA/SLO & reporting of business impact

Service Offering Model (2 of 3) – PaaS Characteristics Custom application development framework with hosting environment Driven by custom, flexible, tailored application requirements Consumers use a hosting environment & “code, load & go” Lack of resources & desire to manage base HW; able & willing development staff Application abstraction layers with built-in scalability Usage is normally based on amount of traffic or compute resources used to service custom application

Management Responsibility – PaaS Degree of Management Responsibility Consumers Service Providers Capacity Change Configuration Data protection Compliance Event Problem Security Asset Server OS Application Service Database Network Examples: Google App Engine Microsoft Azure Service Platform SalesForce Force.com

Monitoring PaaS Clouds End User Monitoring Synthetic Transaction Monitoring Custom support with “record-n-playback” Real User Monitoring Inserting an appliance on network edge to measure outgoing requests & responses Placing a virtual appliance in the Cloud (IaaS) URL and Web Service Response monitoring Scalable and highly effective Stack “pings” Application specific instrumentation Application dashboards URL data gathering Metrics gathered through exercising platform APIs & services Vendor specific Service Level Agreement reporting Automatically generated SLA reports Business Service Management reporting Compliance to SLA/SLO & reporting of business impact

Service Offering Model (3 of 3) – IaaS Characteristics Provides raw computing resources on-demand Driven by Cloud Business Drivers, with attention on technical risk & ultimate flexibility Consumers interact with solution similar to before Smooth transition from internal IT to Cloud Private Virtual Data Centers – “comfort zone” Normally web services/ReST based APIs for management & imaging Usage is charged by the time with no upfront cost, based on the amount of compute resources allocated

Management Responsibility – IaaS Degree of Management Responsibility Consumers Service Providers Capacity Change Configuration Data protection Compliance Event Problem Security Asset Server OS Application Service Database Network Examples: Amazon EC2 Rackspace Cloud Servers Terremark Enterprise Cloud

Monitoring IaaS Clouds Monitoring embedded in Images Instance landscape too dynamic for polling Dynamic registration of agents Existence of server instance is unpredictable Graceful shutdown Expected & unexpected states Support for proprietary Cloud Service Provider APIs www.unifiedmonitoring.com Automation and integration with Cloud APIs Instantiation and aggregation Benchmarking instances Resource “Available to Promise” versus percentage utilization SaaS/PaaS monitoring techniques also apply to IaaS

Why Benchmarks? Traditional metrics are based on percentage consumption of a fixed resource. Cloud Computing metrics are based on metering usage or consumption of an elastic resource Traditional IT UTILIZATION Percentage CPU Utilization Percentage memory Cloud Metrics AVAILABLE TO PROMISE Available to Computing Power (Latency to calculate Fibonacci sequence) Memory access speed Synthetic transaction-based RATES File I/O Context switching USAGE METERING Amount of data written to disk Consumption of CPU cycles

Top 10 Obstacles to and Opportunities for Growth of Cloud Computing Above the Clouds: A Berkeley View of Cloud Computing Availability of service Data transfer bottlenecks Data lock-in Data confidentiality and auditability Performance unpredictability Scalable storage Bugs in large distributed systems Scaling quickly Reputation fate sharing Software licensing http://www.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-28.pdf

Nimsoft is ready, willing & able to help you with your cloud initiatives. Q & A Thank You! [email_address]

Making Sense Of Cloud Computing - by Mark Rivington

More Related Content

What's hot (20)

Similar to Making Sense Of Cloud Computing - by Mark Rivington (20)

More from CA Nimsoft (20)

Recently uploaded (20)

Making Sense Of Cloud Computing - by Mark Rivington