In any industry, customers running critical applications are typically slow to move to something new. Whether it is a new technology, platform, application, or service provider, people tend to be comfortable with the status quo and it is human nature to try to avoid a “CNN moment” or a resume generating bad decision resulting from implementing something new ->thus, people tend to avoid making changes to their critical workload environment. However, as new solutions or technologies become available and mature, the “I don’t want to be first” mantra is eventually followed by “Oh-oh- looks like I am last” when the realization you are falling behind your competition (who have already adopted the new technology) starts to set in. Finding that point in time when a new technology- solution-product has reached an adequate state of maturity to meet your particular needs and requirements is paramount when considering changes to a critical workload environment.
My brush with being a part of a “CNN moment” was from my past life in the Telco world (circa early 2000’s)- my former employers “new technology” was deployed by an Interconnect provider who was connected via traditional signaling TDM links (56/64Kbps links) to a large Regional Bell Operating Company (RBOC) who was running “legacy technology”. Signaling traffic got kinda high (>70%) late one December night and the legacy gear went into congestion and then went belly-up at the RBOC. Massive cascading outages ensued- 10 counties in Texas could not make phone calls, the President’s plane was in route to Texas- they thought it was a terrorist attack and diverted the plane—not a good day for anyone involved. Eventually, the FCC investigation found that the root cause of the outage was that the legacy gear memory buffers were misconfigured- more than likely long ago when the equipment was first installed. Our gear? -- well, we were newer and had enough onboard memory so we statically give each buffer enough memory (and then some) so that the misconfiguration issue was actually impossible on our gear. Moral- just because you have been operating a piece of equipment for a long time and are comfortable with it that may not mean that it is “best of class” and what you should be operating today.
So back to the topic at hand……………..Many companies who are running critical applications on UNIX in their data centers are facing the decision of staying on UNIX or moving to x86 over the next few years. While some companies are being forced to move due to end of support issues and some companies will choose to stay on UNIX for a variety of factors, most companies over the next 3-5 years will choose x86 as the technology of choice for running their critical data center UNIX-based applications.
Gartner- By 2017, 65% of applications running on proprietary versions of Unix in 2012 will have been migrated to x86 (primarily on Linux), making IT planning essential.(*)
Commodity hardware is a term frequently used when talking about x86 architectures- while that may be true at the chipset level- it is certainly not true at the system level and customers should be aware of the large differences in manageability and availability of the different hardware server choices out there. The Cisco Unified Computing System (UCS) provides an easy to manage -- converged x86 design that is being rapidly adopted by the industry. Not familiar with UCS? -- Take a look at the video below
Once the target x86 platform has been decided upon, a harder decision actually has to be made-> figuring out who is going to be responsible for getting from Point A (where you are today) to Point B (where you want to be). Typically this involves a series of relatively complex and large migrations of applications and databases which must be done in a very short maintenance window(s) or with zero downtime. Deciding on who is going to lead your migration is critical to ensure a smooth migration and that your requirements will be met.
Did you know Cisco Advanced Services can lead and execute the migration of your SAP, Oracle and even custom applications and databases to Cisco UCS? Take a look at what one of our customers has to say and check out the blogs from Cisco’s Timothy Stack on this topic as well.
Getting started in your x86 migration path usually requires a business case to be made. Cisco Advanced Services can work with you to provide a no-charge detailed TCO analysis to show the potential savings and help you begin first step towards moving your critical UNIX workloads to UCS.
Please visit our RISC/UNIX Services and Solutions site for more information.
*Gartner, The Future of Unix: Hazy and Overcast, So Reach for the Umbrella, George J. Weiss, et al, April 20, 2012.