Slashdot Mirror
Adopt the Cloud, Kill Your IT Career
snydeq writes "IT professionals jumping into the cloud with both feet beware: It's irresponsible to think that just because you push a problem outside your office, it ceases to be your problem. It's not just the possibility of empty promises and integration issues that dog the cloud decision; it's also the upgrade to the new devil, the one you don't know. You might be eager to relinquish responsibility of a cranky infrastructure component and push the headaches to a cloud vendor, but in reality you aren't doing that at all. Instead, you're adding another avenue for the blame to follow. The end result of a catastrophic failure or data loss event is exactly the same whether you own the service or contract it out.'"
no one even knows what the cloud is. It's everything, it's nothing, it's cheaper, it's not.
run your IT shop like everything else, with common sense. Can external hosting work sometimes? sure, if so, do it and stop worrying about it.
Since you "know computers" it will still be your problem.
If I had a dime for every time I got blamed or was asked to fix something that was clearly outside of my sphere of influence...
well, I probably wouldn't be reading slashdot right now.
I think that if we adopt the cloud model for our internal networks (i.e. a private cloud) that would help improve manageability of the network.
Rather then outsource to someone else's cloud, create your own.
I sort of agree with the blurb that started this thread.
Instead of being a skilled professional with power to change things and work on a problem, when you go to the cloud you demote yourself to a gopher who can only make complaints via a phone call when things don't work.
Aside from making yourself much more dispensable ( "Well, gee, *I* can call and complain too") you get the frustration of feeling powerless. At least with your own systems you can go in, take readings and try things.
I guess "cloud" at this point means, "Running your programs on a computer with a network connection."
Palm trees and 8
This has nothing specifically to do with "the cloud" at all. It's the same problem you have when you outsource anything -- the company you hired might not provide the quality you were expecting.
Can we please stop the re-hash of old ideas with buzzwords attached? This is a site for engineers, not MBA idiots.
There's no -1 for "I don't get it."
Campus decided to outsource our e-mail to Microsoft BOPS, rather than just do Exchange (or something else) on campus. Problem was that doesn't mean that suddenly campus IT just gets to say "e-mail isn't our problem, call MS!" No, rather IT still hast o do front line support but now when there's a problem you have to call someone else, get the runaround, finger pointing, slow response, and so on.
Net result? We now have an Exchange server on campus and do e-mail that way.
It isn't like outsourcing something magically makes all problems go away, particularly user problems. So you still end up needing support for that, but then you get to deal with another layer of support, one that doesn't really give a shit if your stuff works or not.
Basically people need to STFU about the "cloud" and realize that it is what it always has been: outsourcing and evaluate if it makes sense on those merits. Basically outsourcing is a reasonable idea if you are too small to do something yourself, or if someone does a much better job because they are specialized at it. If neither of those are true, probably best not to outsource.
You save the hassles of maintaining a file server, daily backups, etc. Also gives more features as in the ability of sharing some docs with third parties for example.
Of course when the Feds seize the server because some users have been sharing their music and movies with third parties, you're screwed.
http://dilbert.com/strips/comic/2011-01-07/
Upward mobility is a slippery slope - the higher you climb the more you show your ass.
Moving to the cloud is easier, which is why we keep considering it. It is easier to off load the work onto some cloud operator who is supposed to do it better and possibly cheaper, or at least it LOOKS easier. No more dealing with backup tapes, No more dealing with software licenses and the like, just pay your vendor of choice copy all your data onto the cloud and start tossing hardware and the people that managed it out the door.
Problem here is that doing this job right, on a budget, and on time is FAR from easy. Plus, it is going to be very difficult to verify that your vendor is actually doing the job correctly, considering that the hardware isn't accessible, being located in some server room some distance away. Who knows if they actually do backups of anything, much less actually do off site storage of recovery media. My guess is that as competition in this area heats up, prices will fall with quality falling too. Costs will be trimmed by eliminating skilled labor and without skilled labor the whole house of cards will fall.
Seems to me that the cloud may be a short term gain for most, but in the long run, dumping your infrastructure and the people that go with it is going to bite you eventually, unless the business is very small.
Finally, the biggest messes I've had to clean up had very little to do with a hardware failure or some loss of data. The worst messes I've seen where caused by some administrative error.... Replacing the wrong disk in the RAID, causing the total data loss or not thinking though a command before hitting enter. I don't see how being on a cloud will fix this kind of thing.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
They didn't involve my office in the process at all. They knew they wanted to dump their big ERP for something else, but they chose a cloud based SaaS solution and we warned them that it was probably not a good idea considering their size. Now we get tech support calls almost every day complaining that the SaaS website is frozen, and all we can do is shrug and call the SaaS company's support line because we have no control over it. My boss didn't want to tell them "I told you so" but...
Occasionally living proof of the Ballmer peak.
the "cloud" is the latest (in a long line) of over used buzz words
are you running a few virtual machines on a couple of midlevel servers? probably not "cloud computing"
are you considering virtualizing a large number of servers to achieve high performance/high availability/infrastructure as a service/or some other "as a service" buzzword? probably "cloud computing"
where your "cloud configuration" exists is another issue. there was an article (Forbes maybe) that pointed out how much less money is required to start/run an "Internet startups." With the "public cloud provider" being Amazon Web Services (i.e. just because you are using the "cloud" doesn't mean you outsourced everything)
remember that "I.T." is about helping a company do whatever it is they do - the need for "I.T. people" (especially in security, virtualization, and developers) is not going away, but if you are a "hardware only" tech, spending your day replacing power supplies and installing new hard drives, you don't have a future in corporate IT departments ...
It ain't what they call you. It's what you answer to. http://mylyceum.us/
I have been shouting this from the rooftops for the past two years. People still just don't get it.
If I were God, wouldn't I protect my churches from acts of me?
http://dilbert.com/dyn/str_strip/000000000/00000000/0000000/100000/60000/0000/400/160498/160498.strip.print.gif
Couldn't find the "mashup" on dilberts site, but still good.
Plus, your cloud vendor is in the IT business like everybody else. Like any IT business (or any business for that matter), they have a bottom line to watch and will choose to provide the minimal acceptable service to maximize profits, or they will charge a lot extra to exceed those standards. But, unlike your business, when they need to make business decisions that lead to cost reductions, you don't necessarily know about it until it is too late.
I have to agree. This summary is, well, crap. Anyone trying to "push problems" to somewhere/someone else rather than resolving the problem shouldn't be working in IT for a start.
This is another "fear the cloud, it eats babies" post, which are becoming more frequent recently. I know I'd never make a decision of how/where to host apps/services purely on one criteria, eg: getting rid of my local headache.
Yet another failure of an IDG article.
Outsourcing is outsourcing, whether it's to India, a contract house, the cloud or your own user base.
Many companies outsource internally all the time. For example, once upon a time, you turned in all your expenses to a $30,000 a year office assistant to compile and post the expenses. However, now they have "saved money" by eliminating this position, buying some fancy (and slow and unusable) third party web application, and instead of having someone inexpensive and familiar with expenses and expense policies doing the job day in and day out, we make the $100,000 a year developers and managers task switch to using this third party program. The net effect? One department gets to save $30k a year, while all the other departments probably have a cumulative loss of $100,000 or more.
If you are not allowed to question your government then the government has answered your question.
True, but the flip side is that I have the ability to keep a one-day non-cloud outage to one day by putting effort into it. I have control over what happens. If it's a non-cloud outage I have no control over how long it'll last. That all depends on the vendor and how much priority they put on fixing things, which leaves me in the unenviable position not of laying on the beach but of constantly being on conference calls having to tell upper management "No, we don't know what happened. No, we don't know when it'll be fixed. No, we can't do anything to speed recovery up.". Which, trust me, upper management does not like one bit.
I have been shouting this from the rooftops for the past two years. People still just don't get it.
That's not cloud hosting, that's just remote hosting, colocation, virtual server, whatever. Cloud services are where there's a pool of resources and you ask for resources as you need them and you are charged as you use them and you don't have to worry about anything other than submitting jobs and waiting for results. And your data is stored on their server and those pay-as-you-go instances can access it.
A lot of stuff is called cloud that has no business being called cloud. It's only remote, or hosted.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
We're looking very seriously at the cloud for all new deployments and likely catching a few existing systems.
Not generic things like email or whatever, but for our own company applications. Cost is a major consideration sure, but honestly the biggest win I'm looking for is being able to specify a deployment in code (XML, whatever) and actually see it executed correctly and timely. The ability to deploy an entire infrastructure with the same ease we currently have of typing "make all" to compile.
Server allocations, network ACL settings, storage needs, all of it. All the stuff that currently takes 3-5 teams (DB operations, Sysadmins, Network Operations, etc, etc) a few weeks or months to do, screw up, screw up again, redo thrice, etc. None of this is particularly fancy or new, it's the same basic requests every time. Yet IT can never, ever deploy anything quickly, accurately, or efficiently.
And it's not just this company's IT. It's most every company's IT department. I know, I know, there's a bazillion reasons why this or that can't happen in whatever way, etc. I don't give a flying fuck about the excuses, by bosses sure as hell don't, at the end of the day NOT ONE is ever valid.
The cloud promises to replace all that repetitive deployment headache with the ability to simply specify what we need in a tidy little XML file and press Go. We're talking about taking a part of our SDLC that previously took weeks or months and doing it in seconds. Accurately. Reliably. Repetitively. Without complaints. Without obstacles. Without lost email. Without fat fingers.
That is why your IT department should be incredibly scared of the cloud. Because you've been doing a shit job for decades and now someone has finally figured out how to literally replace yall with 5 lines of script code.
This isn't a question of outsourcing ("internal" clouds are just fine), this is a question of obsolescence. Most of the human hands in a typical IT department are going to have all the modern relevance of a horse and buggy repairman.
My
Why don't we use hosted servers?
Think about it like this, a [industry] business is based on information, confidentiality and good reputation.
How would you like it if you were involved in a [transaction] with, say, a [business] who mishandled funds related to [your stuff]. Your [representative] puts all his files on a hosted EC3 or Azure server and there's a breach. Those documents could well include confidential communications, financial information, etc., etc., etc., Now your confidential information is out in the wild.
What would that say about the trustworthiness of your [representative] and his/her processes?
We employ multi-layered security to ensure that doesn't happen. We control who has physical access to our VM infrastructure as well as network access. Those who have administrative access are all employees of [business]. They're not employees of a third party who has no stake (other than retaining the revenue stream) in the success of the [business].
And I haven't even touched on the network bandwidth issues -- we have to manage and process huge amounts of data, much of which comes from our customers.. If I send you a couple dozen DVDs, will [your hosting provider] load them up onto the server? No? Then we have to transfer huge datasets of customer data across the internet So then we need to increase the size of our network pipes. What's the latency between [provider's] data centers and Europe? Asia?
We get anywhere from sub-millisecond to 10-15ms latency between our offices and our virtual infrastructure. Unless we move our offices into the [provider's] data centers, we won't get anything close to that.
I could go on and on and on.
Bottom line, hosted servers are great. 95% of our servers are VMs. We just host them on our own virtualization infrastructure. If you're a start up or a small company like [other person], it *may* make sense. For medium and large businesses, not so much.
It's all about fitting the infrastructure to the business model.
No, no, you're not thinking; you're just being logical. --Niels Bohr
The IT business loves its fads. Remember client/server? Remember when green screens were passe and everything had to be rewritten as a GUI? Remember when Novell Networking was all the rage? Remember when IBM's Systems Application Architecture (SAA) was hot stuff? Remember when COBOL then Java was going to be platform-independent and displace all other languages? Remember when everything was going to be outsourced to India, then Brazil? Remember when Unix then .NET was going to rule the world? Remember CompuServe, AOL and Prodigy, each ignoring the coming Internet?
.net are still here.
IT loves its fads, but then it gets tired and moves on to the newest shiny thing. Cloud computing is no different; this fad shall pass. But part of the fad will still be with us; after all, both Unix and
If I used a sig over again, would anyone notice?
If it's only one customer upset enough about the problem then the $49.95 applies.
Even when it's as you've written above - that multiplied by all the customers on that server for instance, the care factor can still be close to zero. I've seen that with hotmail and a DNS configuration typo that put the email out of commission for a medium sized University among other customers, yet it was nearly a week before the problem came up in the queue to be fixed. The hosting provider didn't think it was likely that their customers would jump ship for a single outage like that, thus they expected zero financial repercussion, thus a care factor of close to zero.
Get your company's lawyers involved in negotiating with "cloud" providers. Make sure all the things that the "cloud" provider can screw up result in substantial financial penalties. Lawyers are paid to prepare for contingencies like that. If a "cloud" provider won't agree to enforceable service agreements, price out business interruption insurance coverage for the cloud provider's failure. Now you have a backup plan and costing for it.