Does your server rack look like this?

Does your server rack have something like this in it? Are the nerds explaining how its the next big thing in computing.  Has it been mostly empty for a while? You’ve been had, your money was wasted, and you should fire your IT guys who recommended it to you.   Welcome to a blog devoted to outing bad IT practices that are costing you money, time, performance objectives.

2 Blades 14 more to go!

2 Blades 14 more to go!

This is a Blade Server Chassis (Dell in particular) and this is an example of one with 2 servers installed.  The idea is that you buy stripped down servers that share power supplies, network and storage switches, as well as large fans allowing for super dense deployment of servers with far fewer cables and a single vendor that supports the entire thing.  In some cases it can be a good idea, you spend some money up front and once you fill it up you save on cabling, hassle and redundant components. When you see one like this however (especially one that is two years old or more) you’ve been had. The chassis itself is a box with just fans and power supplies runs a minimum of $30k and after support/networking/storage network and other goodies are added this easily doubles. This is fine if its supporting all 16 half hight servers and reducing cost from fully loading those, but in a case like this your paying two to three times what servers costs, and by the time you get around to filling it out 5 years later the chassis itself will be out of date, causing an expensive rip and replace of not just your servers but ALL of your infrastructure to start this over. I’ve seen one of these loaded with only the equivalent of 3 desktops sold installed and configured for $150,000.

Why do people buy them then abandon them?

  • Unrealistic growth and budget forecasting
  • Budget forecasting that comes in big chucks every 4-5 years
  • They Heard it was the “Enterprise” way of doing this
  • Complete lack of understanding of performance needs to accomplish business objectives
  • Find an excuse to put it on another department’s budget
The FYIG’s take:
If blades are being deployed they should be deployed at least two thirds with plans for being full within a year. Your IT staff should have clear tracking of CPU/Memory/Storage/Networking usage metric’s and how this dense expensive rapid growth tool is going to solve current problems as well as the clearly projected near future.

About FYIG:
 
Fire Your It Guys is a blog devoted to calling out bad IT, and specifically calling out the individuals that implement it. This could be anyone from the lonely desktop tech who creates spaghetti wiring, to the CIO who deploys $100,000 gadgets that functionally do nothing but blink lights at you.  Our team hopes to make this an interactive blog and we will be initially taking email’s and working towards this becoming a user driven website.
Please feel free to send us questions and quick pictures! The format will range from pictures and explanations like this article to short tweets of IT behaving badly.  As IT is projected to continue to loose over 100,000 jobs a year, we would like to start a discussion here between the insiders, the C-suite and the grunts about what are simple indicators to tell the difference between the cream of the crop, and the flops.
– The Consultant
Advertisements

About fireyouritguys

Just a hired gun just trying to set things right.
This entry was posted in Blade Server and tagged . Bookmark the permalink.

4 Responses to Does your server rack look like this?

  1. Warren says:

    my rack at home looks like this

  2. Actually. Dell had an offer to get a free chassis if you bought 2 bladeservers.
    It was a really good deal focused to get companies interested and Dell ran at loss with this deal.

    Not all companies would ever fill up a blade, but some would and eventually give Dell an earning. So in this case it’s possible they bought 2 blades and got the chassis for free 🙂

    • Maiira says:

      Stewart,As a VMware VTSP I am biased I admit but I don’t accept the comment about Central Point of Failure (vMotion deals with this) or the Extra Cost as far as ESXi 4.0 is concerned. I would totally agree with you with regard to MS hyper-v though 🙂

  3. Even in the biggest enterprises I don’t see blades being deployed. Very similar experiences – heavy sales from vendor got the first few in place. Once they were in place the realism of what had been purchased set in. Even buying them fully loaded from the beginning the boxes were a management nightmare not fitting into any existing or conceivable management and talent structure which is critical in the enterprise. Shoving the networking, server hardware administration and systems administration into one box with no clear demarcation as to who owns what and how work is to be done. The only way to do anything is to have network admins on the server console or the server admins making network changes and the datacenter has no idea what to do about anything and can’t find a way to isolate themselves from costly mistakes like decommissioning the wrong server since everything is interconnected!!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s