Thursday, March 31, 2011

Inside cloud computing

Recent study by Yankee Group revealed, by 2014 cloud computing will be $42B market. Much is being talked about cloud computing. Some people believe even today, it is hype, because it uses already established computing technology, others think it is disruptive trend representing the next level in the evolution of the internet. Few years back in Web 2.0 Expo a question was asked “What is cloud computing?”  The answers were captured and uploaded in the YouTube by Joyent, very interesting videoWe learn cloud computing definition from users’ perspective; it delivers services to the consumers in a simplified way. As an user you do not need to know how you are getting the service. But the enterprise adopting cloud computing model need to know the detail about: cloud computing architecture, what are the requirements, how to develop, how to implement, what are all benefits, so on and so forth. Here I will look into the basics of cloud computing from a very high level.

The National Institute of Standards and Technology (NIST) provides somewhat more objective and specific definition as depicted in the following picture:
Let’s focus on Cloud computing characteristics, we see, cloud computing architecture has two distinct domains that enable cloud computing.  They are Infrastructure, which is comprised of Network, Server, Storage, and Software. Infrastructure  which will provide on demand self service capability, resource pooling, rapid elasticity, and measured services.  Network component of infrastructure will provide broad network access. Other domain is the end user, who will use the cloud services.  
Infrastructure – On demand Self service: IT Infrastructure needs to provide secured computing resources 24X7 with five nines availability, and should be resilient to enable cloud services. A virtual platform is ideal to provide computing power.
In the Cloud computing era business resilience has moved us from the sense of reacting to recover from an event to becoming impervious to the event. Business continuity focus upon a defensive resilience posture, it consist of three building blocks – Recovery, Hardening and Redundancy – that are widely recognized as vital ingredient for successful business continuity plans. A defensive posture is useful in protecting the organization and its revenue streams but it does not help the bottom line that will support cloud computing environment and overall business growth.
Three further building blocks that support an offensive resilience posture, which are focused upon improving the organization’s competitive position in the cloud environment – Accessibility, Diversification and Autonomic computing. In practice these building blocks can be used all together or in various combinations depending upon need. For example diversifying operations might allow hardening to be limited other than at sites where critical applications and data reside.
Business Resilience should encompass:  continuity, availability, security, recovery and scalability spanning and supporting six discrete layers to support on demand self service for greater QoS to the end users.
IT Infrastructure – Rapid scaling: Scaling an IT infrastructure means managing the performance and capacities of each component within the infrastructure. The basic objectives of scaling a component/system are to:
  • Increase the capacity or speed of the component.
  • Improve the efficiency of the component/system.
  • Shift or reduce the load on the component.
  • Improve overall customer experience
Increasing the scalability of one component, the result may change the dynamics of the entire site service; thereby it is moving the bottleneck to another component. The scalability of the infrastructure depends on the ability of each component to scale to meet increasing demands. So we need to take a holistic view of the IT Infrastructure that include server, clustering, network, storage, workload, and understand the application environment.
Infrastructure: Resource pooling is more about computing resource management to support 24X7 and 99.999 availability. This can be accomplished by adopting virtualization technology. Server, storage and network virtualization can support computing resource requirement more efficiently, but to have optimum computing resource and cost/performance perspective it is essential to do right provisioning by adopting just-in-time approach. Remember most virtualization solutions are not efficiently provisioned and adopt just-in-case approach.
To enable measured service an utility computing services require. The utility computing system (also referred to as a utility system or simply a utility) is a system that can automatically create and manage multiple utility computing services (utilities services, for short) on a shared infrastructure.  The infrastructure consists of pool of hardware resources, such as servers, storage, and network appliances, as well as software resources, such as operating systems, middleware, and applications.
Utility computing services is the “pay-as-you-go” model.  To implement this model, one needs a flexible way to meter the services and resources being used. The UMI (Universal Management Infrastructure) architecture, designed to provide common functions that are needed by most, if not all, of the utilities in a utility computing system, includes a critical metering function.  The architecture of the metering system is hierarchical and highly flexible.  The metering service architecture should include how UMI's metering service function is used in the context of utility computing services, for collecting and storing metered data, computing service metrics (which are useful to the data-consuming applications), and feeding the metrics to various consumer modules (e.g., for accounting and billing).
Infrastructure: Broad network access can be accomplished by augmenting existing typical resilient network topology, like point to point connections, mesh configuration, ring configuration, various  fiber  transport to support failover, through newer trend like Yankee Group’s Anywhere IT approach.   
 We need to do lot of work to support the cloud computing model at infrastructure level. Eliminating detail foundation work may deliver Cloud computing for the enterprise faster, but at some point need to take a holistic look and assess the total need to support cloud computing. I am creating a detail approach and reference architecture to enable enterprise cloud computing for our clients. Please contact me for the artifacts. Before closing this article I would like you to think … how SOA, Web services, and Cloud computing fit together? 



No comments:

Post a Comment