Archive for April, 2011

An Enabling Architecture for Cloud Services

Today’s post is from Rahul Bakshi, vice president, managed services strategy & solution design 

While many different types of architectures can support cloud computing, architecture can limit a cloud’s capabilities and the therefore the use cases for cloud. Cloud architecture should not be proprietary in its technology so that it limits the applications that can be deployed.  The more agnostic the architecture is to the applications, the better.  For example, some databases, are not designed to be virtualized so dedicated computer resources, at a minimum, are required for this type of application. 

 Open and Secure

Architecture should be designed to offer as much flexibility as possible without sacrificing quality.  It not only has to be accessible to end users via various network connectivity requirements, many solutions require the need to support hybrid connectivity. A closed architecture would prevent one application from communicating with another application. For example, an e-commerce application in a closed system might not be able to communicate with a manufacturing application on dedicated infrastructure required for order fulfillment.

With Cloud computing, the increased dependency on security solutions has risen significantly.  Companies are looking to understand the layers of security the cloud solution offers as well as how that security pertains to their specific environment.  Depending on the business or application requirements, general-use “consumer clouds” will not support the appropriate level of security controls or compliance.  Most lack the ability to identify, prevent and track access, attempted access, and actual intrusions and may not include controls for authorized access.  If the infrastructure lacks the controls to detect, log and perform forensics against an intrusion, it limits the types of use cases.

Redundant and Agnostic

In addition to security, true enterprise-grade for cloud offerings means high performance, scalability, and reliability.  Enterprise cloud solutions must provide appropriate layers of redundancy to support true high availability for the application layer.  Redundancy must be built-in across the infrastructure and associated tools ensuring there are no single points of failure as well as seamless failover for the application(s).  This requires automation and appropriate tooling to prevent any requirement for human interaction.  Further, capacity management and process automation are required to maintain the right levels of availability.    Special automation should move the workload wherever needed to maintain availability. 

Performance transparency and automation

Cloud solutions require tools around monitoring, reporting, and managing now more than ever due to how clouds are architected, shared, and made available for applications.  Organizations need to understand how the resources are being used so they can manage and plan for capacity and growth while having confidence in the current performance and health. 

Automation is also a key component of cloud architectures as automation delivers on the value proposition of services needed with high degree of quality.  Operational automation to provision the services businesses need drives costs down while improving time to market.  Integrating this automation into the overall performance management delivers capacity on demand.