Cloud impact on platforms and applications – a perspective on architecture

By Shripadraj Mujumdar, Cognizant Technology Solutions

Context

Today’s large businesses are heavily characterized by globalization and interconnectedness. Therefore, the key effective trends of importance can be summarized as- Business process improvement and consolidations, supply chains. Reducing enterprise costs and managing change initiatives, increasing the use of analytics. Improving enterprise workforce effectiveness, enable innovation and Targeting customers and markets more effectively. In a nutshell, organizations are looking for shorter lines of communication, deeper relationships between stakeholders, and more shared knowledge and data of the company and business, up and down the ranks and across linked partners.

Similarly, with ever-increasing definitions of the enterprise, boundaries to support above key focus areas in an organization’s set of technologies which are playing a pivotal role can be identified as virtualization and Cloud, Service Oriented Architecture, Web 2.0, mobile technologies, unified communications, business intelligence and document management and storage. The game changer here is specifically “The Cloud,” which provides the necessary model to commoditize the platforms, Infrastructure and applications.

Software application architectures have traversed the path of separation of functions, objects and  layers till separation of concerns in terms of services. The new applications are more open, collaborative, and social, and the main factor for facilitation has been maturity and adoption of standards for interoperability and compliance on feature sets. With Cloud power playing its role, this core definition of software and applications is set to undergo yet another paradigm shift, impacting the direction on how Enterprise IT systems are architected, leveraged and governed.

Impact of Cloud on platforms and architecture

In most of the literature, more often than less, the advantages listed for Cloud Computing put more emphasis on its infrastructure aspect and its economic benefit. The economic change that may come to computing services is in terms of changes in patterns and factors of production and consumption, besides aggregation of supply and demand. Moreover, the ability to make over the counter choice and substitution is also an important factor that will impact overall economics of computing in times to come. The real power of the Cloud model will be realized in terms of how platforms, applications and services will get proliferated on pay-per-use or more creative models. For applications which follow new generation of styles, here are possible considerations for impact that may happen in the architecture definition and process. In the future, off-the-shelf patterns will emerge to internalize many such considerations and we may hopefully even see support available at a platform/framework level.

• Business Architecture: The base-level business processes would be set to benefit more from Cloud-based IT service choices in an open model, thus impacting overall business architecture at the enterprise level, driven by the need and ability to explore opportunities for optimization, enhancement and scaling
• Enterprise Assets Reorganization- Based on the speed of adoption, there would be a trend to consolidate the enterprise assets, and replace and reassemble some of them using Cloud-based trusted services, platforms and infrastructure, thus eliminating redundancy. There will be a boost to such service-based outsourcing to leverage cost advantages and minimize in-house maintenance. The data architecture, again, will have a consolidation phase and should be planned to leverage Cloud-based data services – Data as Service — and segregate data within enterprise boundaries and across a public Cloud, creating data virtualization. Moreover, as more Cloud security standards and specifications evolve, potentially there will be shifts in data organization. The possibility of easy access to big data will drive enterprise-level data models to further levels
• Solution and Application Architecture- The most important impact that is happening due to Cloud is on ability of solutions and applications to pop their head out of the traditional box and utilize on massive computing power and scale. Some of the features described below may apply based on the actual application scenario. However, for large-scale, multi-client, multi-scenario applications, all would make sense in the longer term
• Service Orientation- The Cloud-based applications architecture provides a real home to service orientation of applications. With Cloud-based architectures, there would be out-of-the-box support, service end point abstractions, available for SAAS choice available, hence the new applications’ need to expose service interfaces beyond regular access channels to support virtualized service mash-ups. The applications which operate in hybrid mode may well be thought about as assembled from Cloud and non-Cloud parts of services and data
• Multi-tenancy- Similar to the Cloud environment, this facilitates sharing of computing infrastructure across multiple consumers, Cloud applications, running on top of it need to support multiple tenants to avail economies of scale. There can be flexibility to adopt for specific tenancy models based on consumer preferences, accordingly isolation levels can be worked out
• New Integration Scenarios- While applications and business platforms move to Cloud, their interlinkages will invariably have to shift in a similar way. It gives rise to possible new scenarios in which Cloud-based integration can be performed utilizing possible advantage points to find optimal ways
• Cost of Operation/Consumption- The effect of commoditization will directly create a requirement at applications and related level to support similar model. Thus multiple models may have to be supported with blended price structure, similar to support levels we have in regular service provisioning
• Interoperability- The applications and its services need to follow interoperability models considering wide scenario of usage. As of today, the Cloud platforms by themselves don’t set a great example of portability and interoperation. However, the services hosted on such platforms have more scale of access scenarios and need to follow well-established standards to facilitate mash-up scenarios
• Dynamic & Flexible Definition- Considering pay-per-use model and service orientation, applications will need to support the selection of features and services to be used, and pricing model around those services, in order to be competitive and optimal. This is also in terms of transferring the benefit of granular pricing model at Cloud-level to end consumers
• Extensibility- Many sorts of software today also have some extension features which hook up online for doing certain tasks. Going forward, the application extensibility will not be limited to local scenarios, but to be able to choose background services and features for consumption and addition of features dynamically. This will allow the creation of very powerful applications. However, from an architecture standpoint, provisioning for such extension points have to be well thought of
• Elasticity- Configuration & Self Service- In order to support scalability and load variance, the dynamic elasticity on Cloud needs to be utilized in an optimal way. Applications may have to provision for a self-service model for configuration of load modeling and accordingly provision the Cloud resources. Similarly, a self-service model may have to be extended to encompass feature selection scenarios, and pricing will then depend on such selection
• Parallelism– The batch process framework and scenarios in today’s applications may get redefined going forward, due to massive parallelization capabilities which can come out of Cloud-based infrastructure. This in essence will also have an impact on certain business processes when latency is removed. For generic frameworks and applications supporting a wide range of consumption patterns, this may have to be configurable
• Context Awareness- The applications may have to be context-aware in order to be more usable. While the applications move to few centralized locations from a deployment perspective, the consumptions will increase due to high levels of commoditization and access, and a common base will have to cater to variety of contexts
• Service Discovery and Catalogs- With multiple capabilities expected out of services and parameters, and pricing models they would support, discovery and metadata of services will undergo extensions with reference to description of services. The frameworks which support service discovery and facilitate consumption will undergo a change to support the same in the future
• Monitoring, Operations- The monitoring and operational support on Cloud components in an application would be more complex and limited by facilities exposed by such components. Thus, based on specific need and Cloud infrastructure, the frameworks and architectural elements need to be considered for support
• Metrics, SLAs- The Metrics and SLA as used in on-premise software may have change in their definitions. Those will be aggregated functions of on-premise and Cloud services/components collectively. Thus, beyond functional aspect of catalogs in Cloud services, the other aspects of operational specifications may also need to be noted and can become important aspect while choosing amongst similar components

In conclusion, The Cloud, which used to be one of the abstractions in traditional architecture diagrams, is now set to swallow other neighboring elements. When Cloud reaches its fully adopted level, it will influence application architectures by driving them to be more componentized and service-oriented to support agility, commoditization and choice. Consolidating on these points, my presentation on the “Reverse Impact of Cloud” on May 11, at The Open Group Conference, London, 2011 will cover some of the factors which are important considerations for being Cloud-ready, architectures and present scenarios which depict application integration possibilities with Cloud.

Shripadraj Mujumdar will be presenting on the “Reverse Impact of Cloud on Platforms and Architecture” at The Open Group Conference, London, May 9-13. Join us for best practices and case studies on Enterprise Architecture, Cloud, Security and more, presented by preeminent thought leaders in the industry.

Shripadraj R Mujumdar [Prasanna], Senior Architect at Cognizant Technology Solutions, India, is part of Cognizant’s Global Technology Office and carries with him an extensive experience of 16 years of consulting and architecting strategic technology initiatives in corporate and customer programs. He has been instrumental in the inception of new service offerings through COEs, competency & delivery excellence, and in leadership capacity to create high performance technology/domain-focused teams. Shripadraj has done substantial work in India and abroad for several top-notch global customers in order to provide technology solutions and consulting — solving their critical business issues. Apart from his work interest , he actively participates in the blogosphere and is avid reader of science and philosophy. He holds a degree in engineering and has undergone the corporate education program on business leadership.