Cisco is redefining the integrated infrastructure concept with its Unified Computing System product line, yet adoption rates are not as expected, and questions still linger over whether unified computing is worth the effort to implement.
Goals of integrated infrastructure
Historically, the integrated infrastructure ideology was to create a universal infrastructure between data center components so that they could be managed and treated as a whole–simply put, unifying the hardware and software to create integrated systems. That is where Cisco Systems is heading with UCS, however, Cisco’s UCS proves to be vendor specific, eliminating much of the heterogeneity found in today’s data center.
In other words, the design tenants behind Cisco UCS make it more difficult to mix servers, switches and other third-party components into the fabric. And that is where Cisco differs from the core concept of the integrated systems ideology.
However, proponents of an integrated infrastructure argue that there is no better way to achieve integration of server, storage and networking systems than to stick with a single vendor for hardware, software and support. Staying with one vendor should eliminate incompatibility issues and simplify management. What’s more, the homogeneous approach of Cisco’s UCS should reduce training and support costs, make it easier to define service-level agreements and lower management overhead.
Is UCS worth the fuss?
Of course, Cisco is not the only big player that’s creating unified data center computing platforms. Hewlett-Packard (HP) Co. and IBM are heavily vested in delivering integrated systems to the data center. This begs the question: Why is unified computing so important?
The answer here is surprisingly simple. Unified data centers prove to be a better fit for cloud computing, where loads, demands and scale change almost constantly. Today’s successful cloud-service organizations are growing at phenomenal rates, as anything that can provide instant provisioning and scale proves to be of the utmost value. With a unified computing platform, scale becomes an easy element to control. If more processing power is needed, just add additional blades. The same goes for storage, as well as communications. The unified approach brings both plug-and-play capabilities, as well as object-orientated reusability to the data center, allowing compute needs to be met in real time, while helping to reduce wasted cycles, storage and bandwidth.
Roadblocks to integrated infrastructure adoption
Even though all of those benefits seem to be quite attractive, Cisco’s UCS and other integrated infrastructures have not taken off as expected, and their adoption rate is nowhere near the levels previously predicted.
There are many facets that contribute to this slow adoption, the first being cost. For businesses looking to follow a path of homogony, the data center has to be completely reinvented and redeployed to meet the requirements of a single-vendor environment. The associated costs usually prove to be prohibitively expensive, because in most cases, the move to homogony involves a transition period where existing data center operations cannot be disrupted.
On the other hand, if a new data center is being built, a homogenous solution may prove to have a better return on investment and an overall lower total cost of ownership than using a traditional heterogeneous approach.
Of course, choosing the path to follow for integrated systems is a difficult process–one that involves combining the needs of the end users with the goals of the business. And a little prediction is needed for the types of computing solutions that will be used over the next few years.
Making the decision
As integrated infrastructure technologies continue to evolve, data center managers must decide whether to follow the homogeneous system designs offered by Cisco or pursue the heterogeneous designs in use by many already.
What’s more, the success of Cisco’s UCS will be tempered by similar concepts offered by IBM, HP and others that are now taking advantage of an integrated systems approach and already have established footholds in the unified computing market.
Although the ideology behind UCS may have changed, the core beliefs for integrated systems still remain the same–combine I/O, processing and storage into a singular, manageable entity that is easy to provision and scale.
ABOUT THE AUTHOR: Frank Ohlhorst is an IT journalist who has also served as a network administrator and applications programmer before forming his own computer consulting firm.
This was first published in May 2011