Disaster Recovery (DR), High Availability (HA) and Resilient Architectures (RA) are all rooted in the same basic architecture framework, approach and processes with the same objective to provide continuous service in the event of failure. Disaster Recovery (DR) is the technical reactive recovery of services, an all-or-nothing approach when services are no longer available due to a catastrophic event - such as a data center going off-line. High Availability (HA) is the design that eliminates single points of failure, e.g. dual power supplies in server, etc. and Resilient Architectures (RA) is a proactive approach that encompasses both HA and DR architectures best practices. It also includes process and resource management to maintain and meet business defined SLAs in providing IT services.

Resilience Service Oriented Architecture is based on the resilience best practices that defines a set of service tiers or service classes. These tiers/classes define the characteristics necessary to provide a defined availability for the business.

The highest level of service is an availability objective with the potential to meet 24x7x365, i.e., "Always On" availability. This availability is referred to as Hyper-Resiliency; it requires a substantial investment but is designed to support the most critical and time-sensitive technology services - services which needs to be "Always On" and available to meet today’s business’ global demands, i.e. when business stakeholders demand higher levels of IT availability to meet end-user/ consumers’ uptime expectations of continuous availability…"Always-On". This leads us to the mantra of nonstop business, i.e., "build to survive failure".

Resiliency – Yesterday & Today

Let us take a step back in history and look at traditional (yesterday) vs. today's resiliency concepts. The diagram below depicts this eloquently:

You may note that at its core the definition/concept of resiliency has dramatically changed from a conceptually reactive insurance to a proactive necessity.

Resiliency Service Oriented Architecture

As mentioned above resiliency does not equal Disaster Recovery. Along with IT Operational Process' best practices, it encompasses both HA and DR architectures. A major differentiator is that Resiliency Service Oriented Architecture is basically a proactive approach in providing IT services as opposed to reactive. In other words, Resiliency Service Oriented Architectures are a "build to survive failure" attitude. The "build to survive failure" mantra may instinctually point to infrastructure solutions, however, the processes and applications' capability to take advantage of a resilient infrastructure play a key part in resiliency solutions.

So what are these processes and applications?

In order to achieve its goal of providing a 24x7x365, i.e. "Always On" availability, the solution design relies on some of the following:

• Application coding techniques that work in concert with the high availability features of the infrastructure
• Establishment of structure operational environments supporting the change and validation process
• All components must meet same service tiers
• Application supports horizontal and/or vertical scaling
• Leveraging monitoring and automation for continuous availability
• Implementing formal processes to support stability within the environment
• QA environments must also meet same service tier
• Applications must be extended to a secondary site which can, if required, be leveraged for Disaster Recovery purposes

For example, if your release and/or change management processes are challenged or non-existent, then no matter how available the supporting infrastructure and workloads are, bad code released into the enterprise will certainly impact availability of the applications/workloads. And, as far as applications go, if they cannot work in concert with the infrastructures available features, then the investment that went into building a resilient infrastructure may not be fully realized.

In my next post I will define "Resiliency Service Tiers / Classes" and how the classes are used to define and guide architecture design.

Bill Cserjes- Principal Architect , Data Center & Enterprise Resiliency, Wipro, Ltd.

Bill has over 30 years of Technology experience across numerous diverse industries, (public & private sectors), including Investment/Retail Banking, Pharmaceuticals/Healthcare, Manufacturing, Electronic Gaming, Cable/Television, Telecom, and Transportation. He has held strategic and tactical roles in delivering Data Center transformations, Highly Resilient IT infrastructure/data center Operations services and Enterprise wide Resilience Architecture solutions. Bill is a Rutgers graduate with a B.S. in Computer Science & Physics.