The following opinion-editorial is written by Ben Muise, Senior Systems Engineer of Continental Resources, Inc.
BEDFORD, Mass.: One of the truths of life is nothing great is ever achieved by playing it safe. Look at Christopher Columbus. It would’ve been easy for him to say “Hey, you know what? That trade route to India around the Cape of Good Hope seems to be working alright. Let’s just stick with it.” Instead, thinking there might be a faster, more direct route to India, he pointed his ships westward. It seemed to work out pretty well for him.
The world of business and technology is facing a similar dilemma. Virtualization has become a new hot-button issue for enterprises that want to improve their computing power while reducing equipment, power and cooling costs and requirements. The safe route – the Cape of Good Hope of IT – is to take the servers you own (and/or buy new ones) from the suppliers you have always purchased from and impose virtualization on them. It certainly works; there have been plenty of success stories about the virtues of virtualization.
Yet just because it works doesn’t mean it’s the best option. That’s what makes Cisco’s Unified Computing System (UCS) so attractive. Rather than creating a retrofit for existing technology, Cisco started from ground zero. Instead of saying, “How can we make virtualization work with what we have?,” Cisco engineers said, “What is the best way to create a virtual server environment?” The result is a complete re-imagining of server technology and how it operates in a system.
Take moving an application from one server to another. In a normal data center environment, moving the entire “personality” of a server – the MAC address for Ethernet and World Wide Name and Port Name for fibre channel – is a time-consuming, complex process. It doesn’t matter whether you’re moving it across the country or across the aisle; it’s likely going to take 90 to 120 hours to bring up that new server.
With Cisco UCS and UCS Manager, you have the ability to insert an abstraction layer between the hardware and the user community in a virtual environment. This abstraction layer makes it possible to move the MAC address, World Wide Name, Port Name, etc., almost as easily as you move the data. You can spin up the new server in hours, or at most, a few weeks.
This ability to move server personalities easily is a huge advantage both for green field virtual data centers and upgrading existing (brown field) ones because it gives enterprises the ability to start small and then increase capacity as your needs grow. You can gain a sense of how much you can stand up on an individual blade, and if it turns out you’ve undersized it, you can increase the available CPU count, memory and I/O port count by moving the personality to a larger blade – creating an almost instantaneous improvement in application performance. For anyone in a financial or manufacturing environment, where there is a constant demand for improved hardware performance, that’s huge.
Another improvement Cisco has made by starting fresh with UCS is a reduction in the amount of cabling required. Typical servers, of course, require eight cables. Blades in a Cisco UCS environment require two cables at most, and often only one. As a result, it’s faster, easier, and less expensive to build or revamp your virtual data center. Fewer cables also mean fewer pieces that can deteriorate, get pulled out accidentally, be mis-wired or otherwise fail. And should a problem arise at 2:00 AM on a Sunday morning – which is the time these things usually occur – it’s that much less cable to trace through to find the source.
A good way to think of the Cisco approach is it’s the difference between designing an airplane and trying to make an automobile fly. Both might get into the air, but it’s pretty much a given that the purpose-built airplane will out-perform the flying car.
With any new technology, there are always questions about whether it will perform as stated. Critics (especially those with their own agenda) like to point out that the traditional server environment has been around a long time, while UCS is new, and by implication, unproven. Yet that’s not the case.
The current technology is already up to what can be considered version 1.3. The kinks have been worked out, and it’s ready for prime time. UCS has proven both its stability and scalability in real-world environments. It’s ready to become a difference-maker for enterprises that want more performance with much greater flexibility in their data centers.
Of course, as anyone in IT knows, having a great product isn’t enough. In the complex environment of the data center, you also need expertise to make it perform to its capabilities. That’s where Continental Resources shines.
We are one of the few integrators that have earned UCS certification from Cisco. We already know the in’s and out’s of the technology, and can help you take advantage of it whether you want to make a gradual changeover or get a fresh start from the beginning.
Like Columbus, Cisco has chosen to defy conventional wisdom by trying a new approach rather than simply following the crowd. And also like Columbus, that bold move has proven itself to be a smart one that is paying huge dividends to UCS customers.
Instead of settling for a retrofit approach, make your move to virtualization – one that incorporates forward-thinking: Cisco UCS. It’s creating a whole new world in data center excellence.
Ben Muise is a Senior Systems Engineer with Continental Resources (www.itsolutions.conres.com), a provider of advanced technology solutions and support to businesses, governments and academia. He has been with the company for 25 years and has spent the last 5 years specializing in server consolidation and virtualization across both RISC- and CISC-based systems. Muise attended Northeastern University and can be reached at [email protected]