EDITOR'S NOTE
In IT Next’s February survey, IT managers ranked data centre transformation as the third most important trend for 2010. But more significantly, they ranked it No. 1 on the utility parameter.

So when it comes to the utility value, IT managers prioritise data centre transformation even over security, which incidentally ranked No. 1 on an overall basis.

That is not without reason. Data centre transformation has become a necessary primary step to support the dynamic IT needs of today’s increasingly 2.0 organisations. In fact, many IT managers would have long kick-started this transformation and pushed it through, but for the sudden onslaught of slowdown that pushed it off their priority lists, for economic considerations known to all.

Also, the silo characteristics of IT organisations—courtesy an unavoidable fallback of the IT evolution process—make it difficult for IT managers to align IT seamlessly to business needs.

Indeed, many data centres still exemplify an air-conditioned quagmire, or in better instances, a warehouse of IT boxes. This range from servers, switches, and of course, intertwined bundles of patch cords. Often, proprietary clusters of solutions are housed, aimed at serving specific requirements of specific departments.

By virtue of their being, these clusters in the data centre still champion the cause of Enterprise 1.0 characteristics, increasingly less relevant in the 2.0 era.

---

An immediate question that comes to mind is while talking about data centre transformation is why 3.x, particularly when enterprises are yet to fully transform into 2.0 entities. While there can be multiple answers to the question, the simplest one will be that a 3.0 data centre will provide a necessary foundation on which a 2.0 enterprise can be built.

Data centre vendors started propagating the transformation to a 3.0 set-up about three years ago, and now that sure signs of an economic rebound have been evidenced, IT managers can very well paddle up the transformation process. Meanwhile, the key elements of transformation have undergone continuous refinement and evolution and hence the term 3.x.

So let’s first look at some of the key elements of data centre (DC) transformation.

Unified architecture: No silo please
This is the first and foremost part of an effective transformation story, as a DC transformation is essentially about moving on to a next-generation networking and computing environment that enables and fosters convergence of different network views and hues.

For such converged view to become a reality, it is important that there is a unified fabric in place that brings in an effective consolidation among diverse network environments such as the Ethernet, InfiniBand, Fibre Channel and iSCSI.

While with 10G, Ethernet holds the promise of addressing the speed requirements of Fibre Channel (FC), it has still fallen short of addressing no-packet-loss and other important characteristics of FC that are essential to requirements of mission-critical applications.

With IEEE Data Centre Bridging standards in place, Ethernet is more robust than before, but the standards on their own don’t ensure a loss-less service, which is key proposition of various unified fabric solutions from the vendors. These solutions have added specific network components to make the lossless feature a part of the data centre architecture.

This unification not only simplifies the network view but also provides substantial savings in terms of operating and management costs for diverse sets of networks.

An important aspect to be noted here is that of convergence, of voice and data. The unified data centre architecture and networking should take into consideration that voice is increasingly moving from TDM based environments to IP, which is already the widely used protocol for data communication.

Going forward, the unified data centre fabric will also be able to provision for consolidation of voice and data networks in the enterprise.

Data centre virtualisation: It’s about money
There are clear and measureable benefits of carrying out virtualisation in a data centre or across data centres in the enterprise.

Given that high redundancy levels of servers and storage resources are common in data centres, a virtualisation strategy can result in immediate benefits in terms of server and storage rationalisations. Application virtualisation also leads to substantial savings in data centre costs.

From the IT managers’ point of view, virtualisation of a data centre brings in a level of abstraction and pooling of IT resources, which comes handy in aligning IT with business objectives.

This is because virtualisation presents a view of IT resources that is more logic-oriented rather than operations-oriented and hence is more amenable to such alignments.

However, the most significant benefit of DC virtualisation is that it sets the basis for data centre evolution in a well architected manner rather than an ad-hoc manner.

At the same time, there is often a risk of falling for an ad-hoc approach to data centre virtualisation itself, which can defeat the very objective of virtualisation and lead to significant hidden costs—and complexities.

Instead, it is okay to have a phased approach to data centre virtualisation, provided that all the phases are well considered parts of a thoughtfully defined transformation strategy, with a clear roadmap.

Managing heat: Air everywhere
Spiralling power and energy costs alone are strong reasons for making ‘green’ a core part of data centre transformation strategy. While virtualisation also has some by-products that add to the greenness of a data centre, standalone green strategies can result in more significant and measureable benefits for an organisation.

For example, IT managers can identify inefficient and energy hogging servers and chart a phase wise plan to replace those by smarter form-factor servers that are more energy efficient.

Too often, data centres operate with more-than-surplus cooling equipment, which ironically is an accepted way of fighting off ‘hot spots’ that get built up in certain zones, largely due to air-flow deficiencies in the data centres.

Experts have pointed out that most of the data centres suffer from the problem of over-feeding, and that the solution lies in identifying the hot spots and facilitative airflow in deficient zones. IT managers can draw measurable benefits also by installing energy monitoring and management software in the data centre.

The use of green cabling solutions, those that lead to cables with thinner diameters or those that employ intelligent patch panels, can greatly improve air flow within the racks as also across the data centre, thereby further reducing the cooling requirements.

In case of new build-outs, ‘green’ has to be grounds-up. The overall design and layout of the data centre has to be such that it facilitates air flow and its management and reduces dependency on artificial cooling.

While these three elements—unified architecture, virtualisation and green—form the core of a data centre transformation strategy, the existing and prevailing best practices would continue to play their roles.

In fact, an effective data centre transformation strategy can work in tandem with several of the existing best practices.

Another important thing that can drive data centre transformation to the extent of making it mandatory, without saying so, is compliance. More particularly, in industries like BFSI and Telecom, it is already a requirement to preserve customer data for a stipulated number of years.

A policy and standards-based approach to data centre transformation will help organisations address such compliance and regulatory requirements.

Also read: How to transform?

---

Editorial panel

• Virendra Deshpande, DGM, Reliance Tech Services
• Jacob Livinstone, Manager—IT, BEC
• S K Paluskar, GM—Marketing & IT, Excel Industries
• Nikunj Karia, AM—IT, NDTV Imagine

Editorial coordination: Deepak Kumar