No one wants to hear that their equipment is at the end of life when budgets are strained and dollars precious. But in the data center, that is exactly what is happening. Companies are being forced to upgrade and replace networking equipment due to some end of life announcements from Cisco http://www.cisco.com/en/US/products/prod_end_of_life.html the ever popular 6500 series switches and components are nearing their end of life. One question looms. With all of the options out there, will traditionally Cisco only shops look to other vendors for less expensive, more feature rich solutions, or will companies bite the bullet and just budget the upgrades where possible? The other alternative is to run on end of life equipment, which is never ideal.
One thing is sure, data center options increase as vendors try to be everything to everyone in a data center. HP has already retired Cisco from their data centers. Does this open the door for HP shops to displace their end of life Cisco gear with 3Com networking components? How about IBM shops replacing Cisco with Juniper components?
While companies struggle to keep updated laptops and desktops going in the 2-3 year average cycle, many have kept networking gear around for significantly longer periods of time as a means of saving money. End of life notices are one thing if you have a chance to budget. This being November, in a lot of cases next year’s budgets are already set. So what do you do?
So the question is….what is the risk? There is lots of Cisco gear on the second hand market, will companies invest in spares before upgrades? Alternatively, will companies continue to utilize the equipment after the EoL for non critical applications and only upgrade what is imminently necessary? CIO’s are going to face some tough decisions and more importantly some tough justifications. CFO’s in many cases are forcing once closed specs to include other manufacturers. Brand names are exactly that..brand names. I’m not picking on any one manufacturer. I’m seeing more and more of this across the board. If you say you can do it – prove it! And further, proving it isn’t enough…prove you are the best at it from a price and performance perspective.
I have seen some crazy matrices! Bits per dollar, packets per second per dollar, VM’s per server per dollar, and the list goes on. I would love to compile a list of all the equations considered. But in these decisions, remember, you want a business partner not a parts vendor. Anyone can sell you parts and support after the sale can be…well….sometimes less than ideal. Decisions can’t be made in a vacuum. This is especially true in a data center which is, and should be, viewed as an ecosystem.
If you want the best evaluation- my suggestion, and what I have done in the past, is put all the “kids on the same playground.” That is, have a vendor round table discussion. I do them all the time with end users. Think of it this way, there are 4 tires on a car. One bad tire decision and the car won’t drive very well and certainly not for long. The key to a great data center design is versatility, agility, longevity, and all within reason. The same holds true for equipment upgrades. For core equipment, you need to understand what the impact is on power, cooling, over-subcription of ports, growth, cabling and end of life. For servers, you must understand what that does to existing equipment, new purchases, power, cooling, and core and/or networking equipment. For facilities, what are you trying to support? If you can’t get everyone on the same page, trust me, you are reading different books. The standards are a start, common sense and true evaluations are the end.
I do feel rather sorry for the middle companies (SI’s, Implementation firms, etc.) as they are likely to take the heated discussions.