Future Proofing Your Enterprise Architecture
In this context, enterprise architecture (EA) has been promoted as essential for the longevity of IT investments. EA is based on sound principles, but its implementation must address not just current business, but also how it will change in the future. If the points of change are planned for, then a future-proof IT infrastructure becomes reality. In this article, we look at the key drivers of change, and how enterprise architectures can be used to address these drivers.
Change in todays IT landscape is driven by three key kinds of factors: business, technological, and environmental.
Lets explore each of these in a little more detail, in terms of their constituent parts, and how they can be addressed:
The main business drivers that lead to changes in the IT landscape are:
Changes to the functionality of systems are perhaps the most common driver for IT change. Functional changes range from minor modifications to existing systems, to development of new systems. When changes are undertaken with a tactical viewpoint as opposed to a strategic viewpoint, the result ranges from hard-to-manage applications to a fragmented landscape, which exists to serve only point requirements.
At an application level, tactical changes come from small, short-term changes to add functionality in an ad-hoc fashion. Eventually, this leads to an unmanageable application or to a duplication of functionality across applications. To mitigate this, it is important to first define what the application is expected to do, and stick to that definition. An application that does one thing well is always better than one that does multiple things badly. Also think in terms of granular functionality as opposed to making a specific change. This will help drive service orientation in the long run.
At a landscape level, tactical changes come from quick-fix applications that eventually stick around longer than they were supposed to. To mitigate the risk of such changes, it is essential to think long-term in the context of the business itself. Most businesses will have a roadmap for at least three to five years if not more, and it is important to align changes to this roadmap so that the IT and business roadmaps are in synergy. The landscape is then driven by this synergized roadmap, and be business-agile.
Consolidation in a business can occur either as a result of a drive to merge divisions together, or a merger or acquisition. When divisions or companies have their own landscapes, as is mostly the case, there is also a duplication of systems. When a consolidation of such landscapes takes place, there is a propensity to retain these systems. This, however, leads to higher maintenance and a potentially fragile and unpredictable integration between systems.
To address this risk, a consolidation must begin with an inventory of systems and their functionality. Where duplication is found, one system of record needs to be identified for such duplicate functionality, and other systems must then be transitioned to use this system instead. Service orientation is an excellent paradigm to be considered when looking at such a scenario. Building business-oriented, coarse-grained service interfaces and choosing the systems that implement these interfaces ensures that the applications that need the service are not tied to the implementing applications, but to the functionality instead.
On the infrastructure side, consolidation, if done along with a good bit of capacity planning, can lead to significant savings by means of server consolidation and virtualization.
While the cost of hardware has rapidly come down, the key component that has remained high is operational expenditure; be it at the level of software where maintenance contributes, or at the level of infrastructure where management contributes.
Infrastructure management costs are influenced significantly by operational aspects like data center power, cooling, etc., as well as by the need to manage more and more systems. Virtualization is playing a big role in bringing down these costs, by means of leveraging lower cost of powerful hardware to create multiple virtual systems on one physical host.
While virtualization as a concept is useful, it still needs to be planned for; unplanned virtualization is a bigger headache than a non-virtualized environment. Similar to the synergy between IT and business roadmaps, the infrastructure roadmap must also be synchronized with these roadmaps to ensure that the infrastructure also scales to meet the IT and business needs.
Software management costs can be reduced by considering a combination of in-house software plus a software as a service (SaaS) model for non-business critical requirements. Microsoft has called this S+S (software plus services) and has followed it up with an interesting set of offerings in this space. Here again, the key is to plan for which parts of the business can actually be run using a SaaS model, and which cannot.
New IT-Driven Business Models
IT has enabled multiple new business models to come into being. eCommerce earlier and now syndicated ad-based revenues on the Web are telling examples. For IT to better adapt to the change it itself is bringing on, it is important for businesses to keep abreast of technology evolution, and predict how business models are likely to change as a result of a specific IT breakthrough.
While these are hard to predict, it is essential to get an early visibility into such changes so you can react quicker. Most new business models have faster time to market as a critical success factor. IT is expected to support this. For this, IT must look to reuse instead of attempting to build from scratch. Reuse in the enterprise space can be realized by using a service oriented architecture (SOA) that can quickly enable new functionality by combining and aggregating instead of creating from scratch.
From the technology angle, the main factors that lead to changes in the IT landscape are:
The evolution of technology at its current rapid pace is a double-edged sword. While it promotes new ways of doing things, it also leads to a rapid obsolescence of earlier technology, and with it, brings other issues such as unavailability of skilled resources, lack of support, etc. However, these alone shouldnt be the factors taken into account when looking at transitioning to new technologies.
Other aspects like what time lines the change entails, how end users are impacted, etc., must also be taken into account. In many cases, a combination works out better: retain existing investments, but also build new applications in newer technologies. In this, ensuring interoperability by using open standards is important, so that integration across technology stacks becomes easier. Cognizance must be taken, however, because the presence of too many disparate technologies in the landscape is not good either, and should be avoided.
Most applications have been targeted at stationary monitors or kiosks, while now the focus has shifted to mobile computing. Computing form factor has hence changed rapidly. New paradigms like Surface from Microsoft, or iPhone from Apple have also resulted to changes in the way a computer is envisaged. It is important to consider these alternate computing form factors when devising software. Software then needs to be agnostic to form factor, but must also be capable of being customized to fit into computers with less computing power than a sixteen quad-core processor server system.
Hardware capability has increased in keeping with Moores Law. The advent of multi-core processors and programming paradigms around grid computing show an even higher potential in the coming years. Future software and infrastructure must then be looked at with the assumption that more powerful hardware is a given. Paradigms like virtualization and SaaS thrive on this assumption. Businesses should look at utilizing these in their long-term plans.
The cost of management of infrastructure and software is a key component of IT spending. Reduction of cost is another driver for change. With greater emphasis on aspects like power and cooling in the context of a data center, application virtualization is a key factor to keep these costs down. On the software side, application virtualization, and keeping software and data on the cloud are being more widely seen as viable alternatives to in-house software deployed on a per-user basis.
The main environmental drivers that lead to changes in the IT landscape are:
Though user friendliness was always a focus of software in letter, in spirit, earlier systems were aimed at software friendly users. The trend now really is towards actually user-friendly software. With the advent of new on-person computing devices and ubiquitous connectivity, software is changing to provide better experience to users by leveraging this connectivity and computing power. In addition, the advent of Web 2.0 and social networking is another big driver of change. In this context, software needs to also be able to take connectivity and capability into account, so RIA and new devices become the supported norm, while potentially the bare-bones browser might take a backseat.
Green IT is being seen as being more and more important, and IT needs to also embrace this reality. In addition to data center focused initiatives like virtualization, etc., other aspects like lower power devices, leveraging more on-client capability, etc. also need to be brought into play.
Earlier, the focus was on making systems work, while now the focus is on making systems work together. Standards like SOAP, and now WS-*, are promoting more and more interoperation between systems. The fallout of this is new systems will need to adhere to open standards for promoting easier interfacing. Older systems need to be refurbished to build standards-compliant facades around their non-standard cores. Another effect is with more and more systems being capable of interoperating, robust integration architecture becomes more important. Concepts like service-bus provide a far better integration paradigm than point-to-point.
The constancy of change is a well-established fact, and the best that can be done is to anticipate change, and be prepared to handle it. Most change in the IT landscape happens is a result of business, technology, or environmental drivers. The use of new technology paradigms like virtualization in the infrastructure space, service orientation and service buses in the software space help address the technology drivers.
Business drivers are best addressed by a long term synergy between IT and business roadmaps to replace tactical moves with strategic ones. Environmental drivers can be addressed by means of envisioning and embracing ongoing changes like people centricity and thrust towards standards-based integration.
Srinivas Rao Bhagavatula is a program director at MindTree.