Carrie Higbie's Blog http://www.datacenterdynamics.com/ME2/Console/XmlSyndication/Display/RSS.asp?xsid=8DA1D508F56546789362AC50F49AB25F Carrie Higbie's Blog Virtualization and de-virtualization - is that the best answer for power on all systems? http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=C5C55A75B23047ADAF229CCB5F0A6271 One thing that is certain in computing is the cyclical nature of data centers. If you think back over the years, we have gone from main frames, to lots of servers, now to virtualization. Some companies that I work with are already de-virtualizing servers due to the growth in applications and their resulting computing needs. Yet others are de-virtualizing applications that do not require 24x7 resources with a plan to power cycle those machines for an increased savings in power and to free up resources for other servers running on the virtualized machine hardware. That leaves us with the realization that the one thing that is constant is change. The vendor wars are fueling the rate of change. But no matter what you do or plan is it ever enough? You can't fix what you can't measure. You can't measure what you can't identify. And you certainly can't measure without some means to measure. Network management software will give you packets passed, bit errors, etc. Intelligent power supplies will let you know consumption (the number you need) as opposed to power delivered with really is only half of what you need. CPU cycles will tell you utilization statistics. All of these together trended over 45 days will give you a better picture of what is really going on with your technology. One mistake companies make is trending for a week or so. This won't catch things like end of month processing and heavy processing days for reports, payroll., etc. We also have some trends and laws in computing. We know for instance that on average the size of an application doubles every 18 months as applications do more. We also know that we can double the speed of a processor about every 12 months now. All that is well and good, if you have the budget to replace your computers every 12 months. But very few shops have that luxury. So here's the tip from those that have hit the glass ceiling of computing. Plan for extra capacity day one. It's fun to see how many servers you can virtualize until you realize there are too many kids on this playground! It is in your best interest to determine which servers would benefit more from power cycling over virtualization. If your servers don’t need to run from 6PM to 6AM, does it make sense to take up resources on a virtualized machine for that server, or is it better to turn it off? Power cycling with automated power strips is easy to achieve and may provide power savings on some machines. In fact, one data center realized a 10% drop in power just by turning off all servers that didn’t need to run 24x7 and this was a relatively small data center. Sometimes the answer is simple and requires just a flip of the switch. Fri, 23 Apr 2010 00:00:00 EST C5C55A75B23047ADAF229CCB5F0A6271 Departmental Budgets are not CTO/CIO budgets - But Companies are Moving in that Direction! http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=2CDFB69517174C50B6F9C4F6F59E6CA8 It started with energy bills and answering the question, how much are we saving? Of course the answer to that was, we don't know because we can't measure our usage. This changed to lots of monitored power options which according to recent surveys about 75% (on average depending on the survey) of data centers have already implemented. Energy bills moved out of facilities and became a line item in under the CIO/CTO's budget. What's next? Well according to several of the end users I work with internationally, the entire data center budget, and I'm not talking about day one build costs. This is predominantly being driven by day two costs including power and maintenance. Much of these moves have been predicated on the premise that one bad decision in one department can have detrimental repercussions on another. Not so bad if is your department driving the train, but terrible if your department is the caboose! At the end of the line, it is the company that suffers. Historically, each department has been free to make their own decisions and do their own thing - forget everyone else. One company that I recently spoke with said that they went back in history over the last 5 years and could have saved roughly 30% of their overall ongoing maintenance budget and 70% in project expenses that were entirely reactionary due to silo decisions. This company's annual IT budget is in excess of 10 million dollar US. Add to this mix, clouds (public and private), colo decisions and consolidation plans..well you get the picture. Now instead of getting a "here go you spend as you like" budget, each department is becoming a function of joint best for the company decisions. This should also lead to more vendor companies learning to play well in the sandbox together. The end result, a better bottom line in DC operations. Mon, 15 Mar 2010 00:00:00 EST 2CDFB69517174C50B6F9C4F6F59E6CA8 Marketing alliance or product endorsement- Know the difference? http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=18756489413341D0A6807658529AFB7B These days it seems that more and more companies are doing joint announcements and forming alliances to go after business. The trick is determining if an announcement is an actual product endorsement exclusive of all other products, and which is a marketing alliance. In order to gain market share companies share leads either formally or in some cases informally. It is a way to share information and bring new customers to the table. Does that mean that one company's product only works with that alliance company's product? In most cases the answer is a resounding NO! Let’s say company's A, B and C form a marketing alliance. This is very popular in data centers today. If you wonder how that alliance is formed, generally it is dollars from both companies that are shared to create a bigger budget for a greater market reach. What happens if you hate one of the products in the alliance? Do you scrap all 3 companies and look elsewhere? I guess if you are emotional about it you may, but it is more important to understand how the process works. Today’s businesses demand open computing. The standards are written so a compliant product can work with another compliant product within the standard. Products are either compliant or compatible to a standard. The difference being that compliant products meet all of the standard requirements, compatible may not meet them all, some may even be borderline proprietary. If you are looking at a certain alliance, in some cases, there may be a proprietary need for the alliance. If you see that, you are going to be stuck with a proprietary solution for the useful life of the technology. If that suits you needs then fine. But beware that you will be stuck with only those vendors for that life-cycle. If one goes belly up or stops making their part, the other parts are going to die with it. Sometimes there is testing, sometimes not. And sometimes companies A, B or C test with several companies! To add to that complexity, the components inside some of the equipment have been even more stringently tested with multitudes of companies during the development phase! And further yet, many times companies A, B or C have additional marketing alliances with competing products from other companies. Let’s face it, it is rare to see a data center with only one technology in it. Are you going to scrap your server company because of an alliance with a switch company? Are you going to scrap your switch company because of an alliance with a certain server company? Are you going to scrap both your servers and switches because of an alliance with a certain CRAC vendor? Of course you aren’t! Unless there is a warranty reason, or you are using a proprietary solution, you shouldn’t be evaluating other solutions based on marketing relationships unless you are unhappy with a current vendor. Use the marketing alliance for what it is, a good way to get information from a multitude of resources and various products. Then see how your current solutions work within that model. No need to throw the baby out with the bathwater as the saying goes! Thu, 25 Feb 2010 00:00:00 EST 18756489413341D0A6807658529AFB7B Is 2010 the year for upgrades? http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=100EBE6622C14CED8CEE5B1173659D13 With the vendor wars heating up, is it smarter to wait and see what “shakes out” or buy now and deal with any fallout later through equipment attrition? I hear that question a lot. Historically, companies have refreshed their electronics every 3 years. The truth is now that companies hold on to expensive switches, servers, etc. until there is a compelling reason to upgrade or the things just die. We are now at a point however where we do have compelling reasons to upgrade. The first reason is power. As equipment evolves, the power efficiency improves. It is important however to evaluate power for the entire communication. For instance, if a switch port is low power but the server NIC is higher power, one must weigh overall communications power to determine the benefit. The same is true for other critical systems such as cooling units. Older models, quite simply, are not as efficient. The second reason to upgrade is technology. We are at one of those rare points in computing cycles where bandwidth is out there to spare (for some anyway). With companies laying people off, technology must fill in the blanks. Technology is a company’s single greatest advantage over the competition; this is even truer in today’s economy. Productivity must be at its peak for a company to remain afloat. Limping along on old technology won’t work. While I will admit that doesn’t mean you have to do a full scale upgrade, a company must look at what is working and what isn’t. If it costs too much to maintain web servers, why not outsource them or upgrade the platforms so that they are more stable? Third, we have great new technology to choose from. Flywheel technology comes to mind, along with air side economizers, etc. While most aren’t of the brand new born yesterday variety, the fact that they are maturing enough to be competitive and reliable means we should start at least thinking about them. Lastly, at least for this blog, is that we have a far more competitive environment than we did even a year ago, especially in electronics. The trick is that we are going to have to get smarter in our evaluations. I have found many end users using new equations for determining cost benefits such as packets passed per unit of currency or per Watt. Vendor based ROI calculations are a good start, but they are not perfect for every scenario. As the electronics wars continue, we will no doubt see some far more advanced means for evaluating them to see who wins our precious dollars. A fun one would be headaches per hour! Thu, 21 Jan 2010 00:00:00 EST 100EBE6622C14CED8CEE5B1173659D13 Twas the Night Before Christmas- the DC Manager's Version http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=B4D46886B3714A4CBA1985E73FE58CB6 Twas the night before Christmas and all through the DC Not a creature was stirring not even a techie The operators were asleep all tanked on eggnog While those of us that write are dreaming up blogs The new year awaits with visions of green Lower TCO's alive in our dreams The vendors align competing for space IBM, Sun, Cisco, HP, and Dell have all  entered the race Our budgets are smaller and scattered about Can one vendor win?  We all have our doubts. The CIO runs the toyshop's accounts But each department wants all the amounts! Power and cooling, are we too cold or too hot? One things for sure, we're not sure what we've got! New messages abound, some good and some bad They all  want us to know, there's gear to be had! Our propellers in motion are constantly spinning Each vendor we meet has visions of winning Our storage, our networks, our servers yes sir The options and promises, our minds are a whirl Our consultants chime in, "I know it all" When in truth, some know nothing at all Like a doctor that can cure ANYONE that's sick Isn't there someone to be our DC St. Nick? With good in their heart and a spirit of truth Leave me one map to wisdom while I still have some youth! Can we all work together,  so I won't be annoyed? Give me a solution before I'm unemployed! Networking's toys has my cabling a mess Servers need cooling, I mean more and not less Security rains on all the parades Cause hackers cause my boss to go into tirades! No cell phones, no email, those days I miss When I could take a day off, and bask in my bliss No crisis, no hack, no "Damnit it's down" No buy this you have to, or you'll look like a clown! New Years is coming, for one thing should you strive What happens in 3 years or 4 years or 5 I'm tired of having to reinvent the wheel Regardless of what your trying to sell Take care of me vendors, and not with hot meals Work with each other!  my DC is real! My business depends on the solutions I choose and business is something I choose not to loose If you are the best at ALL things that you do Then that is something that you must prove Work with me and work with my business that to me would be a great Christmas! Please stop the lies and please stop the FUD The truth is the truth, a dud is a dud If you can't work with me, and with all that I've got Your gear just plain sucks, it's not worth a lot My advise to consultants and vendors is clear I run this DC, and I won't live in fear! My boss won't say, "here's more, please do less" I don't have time to deal with this mess! So for Christmas and New Years please do one thing for me Bring me a solution for 10 years not 3 If you can't be part of the solution I need I won't be sucked into your vision of greed The new year awaits and my budget is set If any of these dollars you want to get Work with my business, work with my plan Step up to the plate and lend me a hand Don't sell me some stuff just  to count the sale Make sure that for me it will work very well I'll be exstatic and full of great cheer And you'll be my partner in the coming new Year Here's holiday  toast to those who work in my DC It's been a pleasure, well mostly,( I'm being PC) Together we'll work in the coming new year And next year enjoy even more holiday chear Our budgets our brains, our goals they will mesh And hopefully we will clean up all of this mess We'll operate smoothly behind firewalls The helpdesk will get far less crisis calls. We'll plan for growth, while we virtualize We'll dedupe our data, be leaner and wize On HIPPA, on LEEDs, on SOX and the like Merry Christmas to all, DC let's say good night! Fri, 18 Dec 2009 00:00:00 EST B4D46886B3714A4CBA1985E73FE58CB6 Not a creature was stirring not even a techie
The operators were asleep all tanked on eggnog
While those of us that write are dreaming up blogs

The new year awaits with visions of green
Lower TCO's alive in our dreams
The vendors align competing for space
IBM, Sun, Cisco, HP, and Dell have all  entered the race

Our budgets are smaller and scattered about
Can one vendor win?  We all have our doubts.
The CIO runs the toyshop's accounts
But each department wants all the amounts!

Power and cooling, are we too cold or too hot?
One things for sure, we're not sure what we've got!
New messages abound, some good and some bad
They all  want us to know, there's gear to be had!

Our propellers in motion are constantly spinning
Each vendor we meet has visions of winning
Our storage, our networks, our servers yes sir
The options and promises, our minds are a whirl

Our consultants chime in, "I know it all"
When in truth, some know nothing at all
Like a doctor that can cure ANYONE that's sick
Isn't there someone to be our DC St. Nick?

With good in their heart and a spirit of truth
Leave me one map to wisdom while I still have some youth!
Can we all work together,  so I won't be annoyed?
Give me a solution before I'm unemployed!

Networking's toys has my cabling a mess
Servers need cooling, I mean more and not less
Security rains on all the parades
Cause hackers cause my boss to go into tirades!

No cell phones, no email, those days I miss
When I could take a day off, and bask in my bliss
No crisis, no hack, no "Damnit it's down"
No buy this you have to, or you'll look like a clown!

New Years is coming, for one thing should you strive
What happens in 3 years or 4 years or 5
I'm tired of having to reinvent the wheel
Regardless of what your trying to sell

Take care of me vendors, and not with hot meals
Work with each other!  my DC is real!
My business depends on the solutions I choose
and business is something I choose not to loose

If you are the best at ALL things that you do
Then that is something that you must prove
Work with me and work with my business
that to me would be a great Christmas!

Please stop the lies and please stop the FUD
The truth is the truth, a dud is a dud
If you can't work with me, and with all that I've got
Your gear just plain sucks, it's not worth a lot

My advise to consultants and vendors is clear
I run this DC, and I won't live in fear!
My boss won't say, "here's more, please do less"
I don't have time to deal with this mess!

So for Christmas and New Years please do one thing for me
Bring me a solution for 10 years not 3
If you can't be part of the solution I need
I won't be sucked into your vision of greed

The new year awaits and my budget is set
If any of these dollars you want to get
Work with my business, work with my plan
Step up to the plate and lend me a hand

Don't sell me some stuff just  to count the sale
Make sure that for me it will work very well
I'll be exstatic and full of great cheer
And you'll be my partner in the coming new Year

Here's holiday  toast to those who work in my DC
It's been a pleasure, well mostly,( I'm being PC)
Together we'll work in the coming new year
And next year enjoy even more holiday chear

Our budgets our brains, our goals they will mesh
And hopefully we will clean up all of this mess
We'll operate smoothly behind firewalls
The helpdesk will get far less crisis calls.

We'll plan for growth, while we virtualize
We'll dedupe our data, be leaner and wize
On HIPPA, on LEEDs, on SOX and the like
Merry Christmas to all, DC let's say good night!

]]> Green Grid to publish a new Design Guide http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=958A095E94B44294A9F94D2B1B68038D If we add standards, codes, and guides from manufacturers and now the green grid, there are literally hundreds of opinions on how we should design a generic data center. While I agree that it is nice to have a guide, the tricky part is that it isn't always easy to find designers that actually read them. It's even more difficult when vendors set marketing messages around a small item in one of them and promote it at the expense of all other systems in the data center. To complicate matters just a little more, most of these guides are living breathing documents. In respect to cooling and power, we have seen significant changes over the last several years. Free air, water side economizers, air side economizers, flywheel technology, etc. have all started changing the data center space. The reactionary time between when new technology is introduced, and it is referenced in a guide may keep some people from adopting it. This could be a bad thing for the industry as a whole. I hope everyone reads these guides for guidance, as that is there intent. I also encourage everyone to read the effective date, understand when new versions are coming out and not use these guides as a replacement for standards and acceptable best practices. The data center is NOT a place you want to install minimally compliant, short term ANYTHING! This is particularly true to infrastructure. By Infrastructure, I'm referring to core services that electronics will rely on including space, air, power, cable plant, cabinets, etc. The flux in a data center is active gear. The smartest decision is to plan for the gear you believe will be there. I will admit it can be tricky not knowing what technology will be out there tomorrow, but it is doable. I help companies do this pretty much every day. Overbuilding is wasteful, under building is more expensive over time. As more and more of these guides emerge, temper them with common sense, a knowledge of your business needs and a good healthy self check of the information you are getting from vendors. Your decisions should be made on the best of the best, not the latest thing available today. There is a good chance it won't be there tomorrow. The best data center is an agile one! Take advantage of the guides as they apply to you. There is really no such thing as a one size fits all data center. Fri, 13 Nov 2009 00:00:00 EST 958A095E94B44294A9F94D2B1B68038D Legislating cyber security for business - really? http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=6A417BF4132A4CF3A3F92A6C834EB46D The Washington Post reports that such legislation may be introduced as early as this week by Senate Commerce Committee Chairman John D. Rockefeller IV, D-W.Va., and Sen. Olympia J. Snowe, R-Maine. Thinking back to SOX (Sarbases/Oxley) and the significant costs to companies, what would this do to businesses? I seriously doubt that most businesses could stand another expensive legislated hit to their IT budgets. I would also be concerned that new technology would not be deployed, purchased or even evaluated if a company has to come to a standstill just to evaluate where they are with respect to this legislation and what they would have to do to comply. No company is going to do this with a moving target. Going one step farther, what happens to technology companies? They would need to prove that they comply before anyone would entertain their solutions. What happens if a company buys a new solution and learns later that it doesn't comply with these security measures? Can you get money back from the vendor? Will companies start requiring statements of certification in their RFI's. The far reaching impact of this legislation could be terrible for our industry. What are your thoughts? Mon, 09 Nov 2009 00:00:00 EST 6A417BF4132A4CF3A3F92A6C834EB46D Clouds, new toys, is it really a good thing? http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=D3EA08CBCB89452DA68065B8EAEC010D Of the customers I work with around the globe, I think that cloud computing works for some applications but certainly not all. In any scenario, a risk assessment is needed to determine what will actually benefit from clouds and what will be technology just because it is the latest and greatest most hyped in the news. Data centers are particularly cyclical in nature. Today's latest and greatest is built on yesterday's latest and greatest ...just new and improved. Clouds may not work or even be necessary for many business servers/applications. I think that something else that is missing is control of data. There is a fundamental concern in most companies about turning over control. Private clouds offer some of that control. But one also needs to realize that there is a learning curve with each new technology and if not properly implemented it can do more harm than good. Companies are getting much smarter about how they spend departmentally from an individual department ROI/TCO. But the ROI/TCO models must be real! Lacking is an overall control of all individual department budgets in a data center. Until one technical CIO steps up and operates under one budget, networking will continue to make decisions that can have an adverse affect on port over subscription, land locking hardware placement, etc. This leads to problems in facilities,and for server folks that may need to place higher processing (higher heat) hardware in places that are not beneficial. Software adds to this as applications need continually higher power servers and more storage as applications double in size about every 18 months. Any one department can make purchasing decisions that can cripple the budget of another. I think that with or without clouds today we are leading to a new type of CIO/CTO. The new CIO/CTO is going to have to understand what is really going on and companies need to stop just buying into what a vendor says but rather make their own assessments, look at standards, and at the business case for their own company. I realize we all want new toys; we want to adapt new technology. But in some cases companies don't even know what that new technology will get them or cost them when all budgets are included. Security, applications, servers, networks and facilities will all be impacted either positively or negatively, but if they are not all involved in the pros and cons decisions - better hold on to your wallet! And in all cases, if a company does not have a good grasp on the resources it currently uses, the power associated with them, and the real risk to business for each application, then cloud computing is just throwing another iron in the fire that also needs assessment. Another issue is the way that marketing material is presented. I see many ROI models that tout a savings against money that would not be spent or in some cases should be spent to allow maximum agility and growth for the least amount of money. If the latter is ignored a company may be locked into a single vendor solution that in fact over time may prove to be the worst. Reconfiguring a DC is not an inexpensive endeavor. In many cases that unrealized savings ends up being an expense later. There needs to be a REAL accountability end to end (translated including all budgets!) before any technology is adopted or considered. Companies MUST plan for inevitable changes, growth and future deployments. Short sighted decisions are by far the most costly. It is time data centers start asking "what if we do this in the future?" One other thing that is changing is that companies that have been traditionally a vendor X shop are now evaluating other technologies due to price, performance, support, and options. Vendor X today may be Vendor Y tomorrow. If you don't plan for change, growth, and other equipment, you may end up with unbudgeted, last minute, overly costly expenses. Sat, 31 Oct 2009 00:00:00 EST D3EA08CBCB89452DA68065B8EAEC010D Overhead versus Underfloor Pathways - which is best? http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=5203181AF7524071A3D0325FE3A25DC9 This is one question that I hear repeatedly. Those that have had a giant blob of spaghetti would vote for overhead systems, those that have had visible overhead spaghetti would vote for underfloor. The correct answer is, it depends. Data centers today often deploy a mixture of both. The error that data centers make is randomly making this decision. In a new build, you basically have a blank slate and you can chose whichever suits your needs and preferences. Either system can really get a bum rap based on poor practices from the past (mostly cable abatement and discipline). The trick is to properly design and maintain either system. I was amazed at a conversation that I had recently with a designer. They put in 12" trays (300mm) trays around the perimeter of the room and one row across each cabinet. I asked how the size was picked and he said it was a historical thing. I then asked, "what happens in the higher density areas?" Now here's the shocker. His response, "tough, they have to deal with it- that is all they get." That is like saying you can have your data center but you have to make do with one server. Huh? Again, the key is properly designed and properly maintained. You can't properly design without knowledge of the networking components, where they will be located, types and number of server connections, etc. In an existing space, the answer is not as easy. Randomly deciding on one or another can be detrimentally costly. For overhead systems, you need to consult a structural engineer to be sure the roof can support the weight. You may need to design around things like lighting, fire suppression, etc. There may be multiple layers of tray needed for the densities you want. This may not be possible with lower ceiling heights. Another set of concerns may happen with underfloor systems. First and foremost, is there physically enough space without limiting airflow? This is generally a concern where the floor voids are limited in height. What else is under the raised floor? Electrical and chilling pipes may cause obstructions that need to be worked around. Local also plays a roll as in some locals working over data center cabinets means they need to be shut down. As silly as it sounds, in some areas it is illegal to have someone on a ladder over live equipment. The list of considerations goes on and on, but in short, the key is to evaluate the space, other disciplines within the space, and density needs. Once installed, make sure to carefully maintain the systems so that a neat infrastructure stays that way and doesn't turn into a pasta dish. Thu, 08 Oct 2009 00:00:00 EST 5203181AF7524071A3D0325FE3A25DC9 The data center battles heat up! Watch your budgets! http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=84069B89ED094EC69E3A2D5F06D2142E I keep in touch via blogs, newsletters, stocks and of course end user opinions and one thing that is true is that the competition in the data center space is getting stronger and even a bit bloodier than any other time I can remember. (I've been in this industry for the better part of 30 years). New partnerships both marketing and OEM are heating up this space. The days of saying I'm a "brand X" house are going to be gone shortly, I think. With all this press coverage, CTO's and CIO's are going to start evaluating these loyalties, and in fact I'm seeing just that globally. When you take the most critical sectors, Government, Healthcare, Finance, etc. the data centers take on a very different role than that of some smaller companies. Don't get me wrong we all know that a company can't function without it's data, but when you add the competitive nature and demands required in these sectors, the winner will be the best combination of value, packets passed, warranty, maintenance, etc. There is not a giant curve anymore to learn one technology as opposed to another. If you have a basic knowledge of how it all works, then the rest is just a command change away and all have the wonder help ? key built in. I was at a show in Sydney recently and was browsing a Juper for Dummies book in the Juniper booth. Pretty straight forward and would be easy to pick up if you are familiar with CLI. Getting past that, the other equations start becoming the deciding factor. I have seen several shops change from their status quo in recent months. When you consider that the standards are written so that products are interoperable, that changes things even more. You don't have to do a wholesale change of equipment like you did in the past when all the equipment ran their own protocols. Today, no gateway is needed. So evaluating technology and a company takes on a whole new meaning. You won't be buying "brand X" just because you have always bought "brand X" (as an example). There are several products on the market for any application. Some are, of course, going to provide better service than others. This is not to say that all of the lower priced products are inferior or that all of the higher cost products are superior. There are several differentiators that should come into your decision factors. These are not based on MTBF (Mean Time Before Failure) or MTTR (Mean Time to Repair) which are basic considerations that should be examined closely. One caveat on the above, is that you must examine the MTTR and MTBF for each and every component, not just the overall chassis or base system. The cards that go into systems may not carry the same ratings. Other factors that do not always come into play include: Value added services Education and Transfer of Knowledge Vendor support Lead times References (not from the vendor) Warranty periods and warranty coverage Training provided to the installer or reseller you select to do your installation Interoperability Day 2 services Money spent on R&D Also look at the number of patches released in the last several years both for security and operations. The more you have to patch, the more time it will take to support the technology. Some companies are swaying away from the most standardized products simply because of the size of the targets on their backs from hackers, or are at least layering on other solutions as a safeguard. Each of the above will make your life easier down the road. Who knows, you might even be able to take a vacation without your cell phone or pager! A vendor scorecard should be developed when you are doing your product evaluations and each of the above should be included with a weighting factor. Weigh each of the above based on the need for your organization. Thu, 20 Aug 2009 00:00:00 EST 84069B89ED094EC69E3A2D5F06D2142E