Slashdot Mirror
US Government Begins Largest IT Consolidation in History
miller60 writes "Saying 1,100 data centers is too many, the federal government has begun what looms as the largest IT consolidation in history. Federal CIO Vivek Kundra has directed federal agencies to inventory their assets by April 30 and prepare a plan to reduce the number of servers and data centers, with a focus on slashing energy costs (full memo). Kundra says some applications may be shifted to cloud computing platforms customized for government use."
I predict a rash of job openings that you can get hired for provided you can spell "Cloud Computing"
[signature]
Dinner is served! Please approach the money trough in an orderly line...
See, the last time we upgraded we put everything on eleven hundred windows 95 machines with 1 gig hard drives. That did pretty good for a spell, all things considered. Now we're thinking about one of them pointy computers... whaddya call em? Blade servers? Yeah, we hear good things about those.
This being a government IT project, I predict it will take 5 years longer than planned, cost 10x the initial budget, and still never really work quite right.
If Virginia's IT overhaul is any indication, this is going to be a slow-motion cluster of a mess for the next 10-20 years
Every business I've ever worked for has had that one dusty 8086 off in a corner. It would run a single batch file every few hours. No one would touch it, because no one knew what it did-- just that whatever it did do was mission critical.
Thus, the US government should just consolidate everything down to a single batch program run by a 8086. I'm sure there's a spare closet in the White House or something they can use as a server room.
UTF-8: There and Back Again
And if you like the way your bank is not liable for identity theft, you'll just love the upcoming government data-filled Cloud!
Caveat Utilitor
Oh, Vivek, what brilliant thing will you think of next? How much energy will it take to replace all of these server farms? How much energy will be required for the taxpayers to earn the money necessary to pay for it? What about security concerns of consolidating all of this data?
I think Vivek wants to make himself look useful after being exposed as a fraud by John C. Dvorak. http://www.dvorak.org/blog/2009/08/12/special-report-is-us-chief-information-officer-cio-vivek-kundra-a-phony/
I am not quite sure what you are talking about. Because everyone has access to yesteryear's supercomputer on their desktop, there is no reason whatever to go back to a 1960s outsourcing model. If you want to distribute load over your machines, go ahead! But why do it over someone else's?
If you think this is going to reduce IT expenditure requirements, you have barely worked a minute in IT. When you outsource, you are simply paying someone else to do your job, plus profit, plus a gaggle of negotiators in middle management collecting their kickbacks, plus downtime costs because your business is less important to them than your business is to you (if you have enterprise e-mail and it has been down more than, say, GMail, you have done something very wrong)...
Didn't read the article, but my experience with government entities is that they receive a specific value of funding each year to spend on gear, training, energy costs, etc.
The nature of the funding goes that if you don't use all of it this year, you get a reduced amount next year. Now this may seem logical -- it may seem like a policy that governs spending. Instead what it is is a policy that drives UNNECESSARY SPENDING.
The places I have been were frugal but appropriate in their spending throughout the year. As the funding for the year would approach a close (in October), all-of-a-sudden the leadership would start spending money like crazy because they had a large surplus. Money would be spent on things that were not actually necessary; if they were necessary, why not get them at any other time during the year?
In several cases, seeing this strange frenzy of spending I would ask the leadership what was going on. They explained the 'use it or lose it' policy and that in order to maintain the funding they got this year, for next year, they *must* spend it all. I was in conflict because I was taught integrity/honesty and there is no integrity in spending up dads helpful money on worthless junk so as to appear that you still have 'need'.
------
The reason I bring this up is because I am curious if the units that will save money via IT consolidation will actually save us money or if they will be (by obvious standing procedure) driven to spend it in pointless/needless ways.
Discuss? Anyone else experience this?
Now just wait for a data center to be scheduled to close in some Congressman's home district and see how big of a block is put into place.
III.IIVIVIXIIVIVIIIVVIIIIXVIIIXIIIIIIIIVIIIIVVIII
I have no problem with the CONCEPT of consolidation, but Virginia's IT outsourcing/consolidation project to Northrup Grumman happened on Kundra's watch. It is an unmitigated disaster.
Years into it, there's not even a complete inventory of the systems that NG is supposed to be managing for the Commonwealth, and at least as of a few months ago, NG couldn't even produce an invoice for the Commonwealth to pay that had more than six or eight line items on it.
I sat through a special meeting of the House Committee on Science and Technology on the issue a few months ago, and the legislature is NOT happy about the situation. Privately, you will hear from them words like "gross negligence" to "I'm convinced it's corruption". The Delegates who engineered the legislation enabling the IT outsourcing are especially pissed.
No disrespect to Kundra, but I don't think he's the right guy to oversee it.
I agree with the government's effort to consolidate, because you can take advantage of cheaper per-gigabyte costs and have more robust backup, recovery, disaster recovery, and redundancy solutions when you're using enterprise equipment in large data centers. I think the government has a lot to gain from consolidation in this manner.
However, I don't see that they'd have much to gain by outsourcing. Government data, by nature, is quite a bit more sensitive than just about any private company's data. The kind of security the government needs is not going to come cheaply, and it's arguable that any private company is really capable of providing it (although they say they are). Even if they can provide it, it's doubtful they can do it cheaper than the government could. For people in need of true commodity services like web hosting, outsourcing makes sense because it can be done far cheaper that way. For people in need of large-scale custom solutions, like the government, keeping it in-house is going to tend to be both more secure and less expensive.
If you think this is going to reduce IT expenditure requirements, you have barely worked a minute in IT. When you outsource, you are simply paying someone else to do your job, plus profit, plus a gaggle of negotiators in middle management collecting their kickbacks, plus downtime costs because your business is less important to them than your business is to you (if you have enterprise e-mail and it has been down more than, say, GMail, you have done something very wrong)...
I've worked in IT for... a few years. And I agree with the GP.
See, the thing is that while huge organisations will continue to require significant IT infrastructure (either managed inhouse or managed by an outside firm), huge organisations do not provide the majority of jobs in this world. The great majority of jobs are provided by SMBs. The really small SMBs have been outsourcing their IT for years - though "outsourcing their IT" probably translates to "get Dave's son to do it, he knows about computers".
Slightly larger SMBs have been outsourcing their IT to some little company who thought they could earn easy money doing installation and support. Look in the yellow pages, you'll find hundreds of little companies offering services like this. Few of these little outsourcing companies are making serious money - there's simply too much competition in the market.
Larger still SMBs (think medium rather than small, 40-200 employees) may have historically had a full-time IT person. But today there are dozens of companies offering outsourced Exchange, or you can sign up for Google for Domains and the price is so cheap that there is no way a single full-time IT person (even if you ignore their salary) can compete economically - never mind offering four or five nines uptime and spam filtering which doesn't leave people crying. Meanwhile, the cost of a single desktop PC is now so low that it's cheaper to have a spares cupboard containing enough spare PCs to re-equip an entire team at a moment's notice than it is to keep someone on staff to maintain them. Sure, they won't be particularly elegantly managed (there may not be a domain, antivirus may be totally forgotten about, they certainly won't have a standardised build) but let's be honest here - how many non-techies ever display any sign of caring about any of that? And business-specific niche software is frequently sold with a support contract anyway.
Seriously - while anyone who takes careers advice from a stranger on /. probably needs their brains looking at, I'd say if you want steady employment with minimal risk of finding that not only are you redundant from your current post, supply and demand has made you worth considerably less since you last were jobhunting - get yourself a job in the public sector or get the hell out of IT.
I've actually done a lot of smaller server consolidation projects. In most cases, the results are great...those lonely database and file servers that get hits 5 or 6 times a day are all combined into one big box that actually uses all the hardware capacity.
The biggest problems I've seen with VMs are the project managers who treat it as magic, never-ending capacity. The new favorite phrase in IT project management circles seems to be, "Oh, we'll just build a VM for it." Problem is, unless someone else is hosting your data center, you can't just call up and order more capacity without paying for more hardware.
Second-biggest with a consolidation like this is incomplete requirements. Lowest-bidder contractors are not going to do a good job of gathering every single requirement...even high-bidder contractors have problems with this. And the problem is that the more they miss, the worse the fallout. A certain large company I used to work for found this out the hard way moving their inhouse data center to one of the big IT services companies. I'm a systems guy, and had all my stuff well documented. Others were pissed off they were losing their jobs and intentionally withheld information...the contractors didn't follow up, and a lot of last minute scrambling had to be done to complete the migration.
Third problem for a government IT consolidation? Some huge services company like Accenture or IBM is going to win the bid and staff the project with dumbasses they pulled off the street in order to maximize profits. (Yes, this happened in my case in point #2 above...the sales staff presented the A Squad and swapped them out as soon as the contract was signed.) Not that government employees are rockstars, but they at least have a vested interest in keeping the data safe. IBM will probably win the contract too, given their involvement with government systems already. IBM has been so India-happy over the last ten years that I wouldn't be surprised if a lot of the (non security critical) work ends up there.
Just like PMs treat VMs as magic hardware, CIOs treat outsourcers as magic black boxes that flawlessly run their IT operations. Unfortunately, the reality is not as sunny beneath the surface!