You’ve heard the rhetoric: Cloud computing changes everything. The days of the enterprise data centre and the small business computer room are numbered. At some point in the not too distant future, you’ll be switching off the last server, turning out the lights, and looking for a career change. The sad thing is that some people …
It's the old integration chestnut
Cloud or no cloud, it's the old integration chestnut all over again:
PHB: We're going to outsource service X because it will cost us less to run it via a third party as we'll have fewer systems to support.
IT: How do we integrate the outsourced service into our current processes?
PHB: We will integrate it with existing in-house infrastructure using state of the art SOA technology!
IT: So we need to upgrade our well-known predictable in-house systems to be able to integrate with unknown unpredictable external systems and we need to connect them over an unreliable and unpredictable network known as the internet?
PHB: That's fine, we're hiring highly skilled consultants to do the job.
Not that I've been here before, you understand...
Death of this, death of that...
In the 20 odd years I've been bluffing my way in IT, the death of <insert platform/language/whatever> has been a recurring theme - thin clients will kill desktops, Java will kill platform specific programming, clouds will kill data centers.
No, they won't - at least, not yet. In my very limited experiences, what I've seen is that new technologies get incorporated in an additional scenario - the new technology gets deployed alongside the older technology (I refuse to call any technology under my own age as old, so Cobol is still cutting edge, okay?), and we get a synergy, or at least, an expanded total.
Maybe certain technologies do get largely superceded - but I still see impact printers working alongside daisy printers, I still have various old CRT monitors lying around because I need them for my work (don't ask), and so IM(very)HO, there will be a need for some reskilling, but the old techniques will still be required.
I'm reminded of a crap joke...
A man has cancer, and just before he dies, his relatives have him preserved in liquid helium.
From his perception, he just wakes up, and wonders where he is, and a guy comes in and says "Amazing, we didn't think it was going to work, but we're desparate, and spent all our resources to revive you because John, we need you - it's the year 9999, and we understand that you're a COBOL programmer..."
Agree with your ideas but again, the terminology is meaningless.
There is nothing different being done if you're talking about a 'cloud' of the private variety than what many business already do with virtualisation and resource pooling / load balancing of virtualised hosts.
Not sure how I feel about the sponsored by Microsoft by line in your paper. Not because it's Microsoft, because it's a company with a vested interest in what you are discussing. Would be dubious if it were AMD/Intel and so on as well.
Whats a cloud?
Try explaining to 5000+ people in a corporate culture that their data is somewhere in a cloud in another country over which they have no control... and anyways, many parts of the world do not have adequate telco infrastructure to go for cloud architecture.
It gets even better...
...try explaining all of that to them right after the local road construction crew accidentally cuts the fiber trunk.
I figure it'd be about five minutes between that explanation (no matter what explanation) and the IT director getting sacked.
"At some point in the not too distant future, you’ll be switching off the last server, turning out the lights, and looking for a career change."
And like ... the cloud doesn't run on actual servers?
Surely the cloud provides just as many (probably more) jobs, just centralised elsewhere?
the cloud actually introduces additional redundancy - just the same way that any traditional outsourcing contract does. Let's say you go out and hire a company to provide X at Y price:
First off, someone has to manage that relationship. Secondly, X will not necessarily include everything a client might ever need or want. In many cases, particularly with Clouds, it's not even enough to get you through a normal business day... since part of that low Low LOW price is that, for the most part, the service is either up and functioning or down and broken (A or B) - if an Admin somewhere is having trouble with a recurring calendar appointment, the provider (again IME) isn't contractually obligated to help them out... they only want to talk to SMEs who have already triaged the issue (not too different than the concept of named callers with software vendors). Someone has to play middleman between the users and the cloud vendors, and it's usually a team of someones.
Same old, same old
If outsourcers could just finally convince security and the public at large that their personal details are safe outside of the company and country (or at least obscure the insecurity). Then this cloud thing could let them send most of what is left of IT in the West to far off destinations (where the BPOs roam) and the CEO can get another zero in his bonus.
Adding complexity to corporate systems ...
... is contraindicated. Always.
Personally, I'm already taking contracts to fix the problems caused by people buying into so-called "cloud computing" inappropriately.
"Cloud computing" is a marketing term meaning "centralized computing". It was probably inadvertently coined by a middle-manager somewhere who "took a class" once on networking, featuring OSI and/or X.25 concepts, with the lower level protocols shown as a "cloud" to shield him/her from the gory details.
Centralized computing has it's place, don't get me wrong. But for the vast majority of users? Not so much. Not with the low cost of CPU, memory & storage in the modern world.
Have an opinion? Sure!
"The Cloud" does indeed mean joblessness. SME Sysadmins won't have a purpose. I see it already as the smaller businesses move away from in-house IT resources to "the Cloud." It's vaguely disconcerting how many operations I know of with less than 100 staff who do everything using Google Apps.
Will the upper-middle-enterprise sysadmin jobs go away entirely? Probably not anytime soon…but they will not need to be expanded beyond their current number either. We’ll move more of this work into “the cloud.” Do more with less people. Large enterprise folk are probably safe: those companies can understand the necessity of keeping a lid on your own gear. They also have the resources to do it.
So where does that leave us? Some jobs will shift into the large datacenters – someone has to keep those clouds fluffed! Others will dry up – probably through retirements – and simply not be replaced. Overall though…we’re probably not talking about /dramatic sysadmin job number declines in IT/ so much as we are /zero to minimal growth of sysadmin jobs in IT for the foreseeable future./
That has a net effect of driving down wages and lowering job security as more youngsters enter the market hoping to be sysadmins than are falling off the stack at the top. (Go into IT they say – there’s jobs there!) What’s more…IT is a global market. We’re subject to the wage pressures of globalisation. There’s the regular pap about “outsourced workers being no good,” but that view is rapidly becoming outdated. Many of these offshore companies have fantastic staff and are even getting their management structures set up to be compatible with western environments.
Cloud services offer “it just works” corporate IT to small businesses. This is appealing as it can remove the cost of an expensive IT worker or consultant. (Or at least reduce that cost.) Medium businesses can now grow larger before they get to the point where an IT body or consultant is required. Globalisation means competition for the jobs moved to those datacenters is highly competitive.
So impacting jobs?
Hell yes it is. I'd love to some actual hard statistics for western countries that show a growth in sysadmin jobs at least high enough to compensate for the massive economy-and-increasing-shareholder-value related job losses as well as the flood of greenhorns entering the market.
Until I see those figures, I'll have nothing but my anecdotal evidence to rely upon. And that evidence says slightly fewer people need sysadmins this year than they did last year…but there are a lot more sysadmins to go around.
I'm Jobs less already.
I refuse to type a title
Yeah, it's just the same old sh1t coming round again. No-one whose been in IT for at least a decade is seriously worried. As long as you're reasonably flexible and know how to fool/play along with management you'll be fine.
I'm an Ops guy, not a Developer
... but I echo Bruno and Eddie's comments. The death of anything in IT is always greatly exaggerated. I heard a rumor recently of a Novell box (Netware 3.x or 4.x vintage so the story goes) still running in one of the environments we support. I'd venture that easily 40% of the Exchange environments we run have a Lotus server in a closet somewhere too. When it comes to Microsoft BPOS (or whatever they're calling it these days) when you add back in the gap filler (we usually call it level 1.5) the solution is usually more expensive than it was before - but people-expense isn't always the point.
For most companies the biggest value of "the cloud" is not cost or staff reduction, but in letting someone else manage all the asset and licensing bull...s...tuff. For publicly traded companies this can also help shape the balance sheet (CapEx becomes OpEx and fixed costs become more variable).
The promise (in my book "hope") of the cloud is that it will also allow your IT folks to concentrate on what is important to your business (optimization of your cloud solutions vs. end-to-end maintenance of your infrastructure). I call this a hope because after working on three BPOS transitions in the last year, just because you call it a cloud doesn't mean it's actually new or different than a traditional hosting/service-provider setup.
It's the Elastic-Computing/Multi-Tenant clouds that are the real clouds in my book, and that's where the game changer is going to be. It's an old story, but I still jaw-drop when I go back and read what SmugMug has done with Amazon EC2 - http://don.blogs.smugmug.com/2008/06/03/skynet-lives-aka-ec2-smugmug/ - *that* is the brave new world of the cloud in my book... and one where existing IT staff will have a challenge to adapt or get left behind, but it's really no different than what we have been through before... the pendulum swings (centralized->distributed->centralized) and we need to adapt/evolve. Some will, some won't but life (and IT) will go on no matter.
I thought this was supposed to save 20% to 30%
Stakeholder: "I thought the architects said was supposed to save 20% to 30%"
Dev: "Only if we build more than one application that uses it because the second and third apps won't need to re-develop the core"
When I was working on a different industry
I used to assist an engineer, in 1988 first it was CAD that was to kill draftmanship, then in 1990 it was engineering software that was to kill the engineer.
Guess what? After 1995 and everybody was using computers for everything, people used CAD software but still didn't knew how to draw neither in paper or computer, and the engineer was having more work than ever because people used to make a mess of the engineering software, because they still sucked at the job.
However this cloud computing thing in my opinion will have and has a lot more impact. In fact, take this as an example: any business of less than 20 people does not have an in house email server any more.
I enjoyed your post but...
I've never known a business of less than 20 (higher than that 'm sure if I really thought about it) that ever ran their own e-mail.
I've always been a big fan of the phrase: "just because you can do something, that doesn't mean you should". At the end of the day companies should be focusing on what is best for their business and their bottom line, not what the latest trend in "[CIO Monthly]" is.
Cloud computing changes everything.
I thought it was Torchwood that was going to change everything....
Flashback to 1987?
Like the posters above, I think we've all seen this repeatedly in the IT industry. Whether it's Unix/Windows/Linux/Apple/Risc/Cisc/Token Ring/Gigabit/Wireless/Whatever... they only bring new technology into the picture.. and it takes years if not decades to evolve the installed application base and infrastructure, that's assuming it's even possible or desirable to make that move.
When asked what skills to develop for a career in IT, how do you answer? " A large tolerance for alcohol, for those vendor funded open bars", "Learning that you'll never be caught up with the technology, and you have to remain in learning mode perpetually", and maybe the most important -- Learn what a UAT* really is, and how and when to apply it.
True IT professionals, the ones who actually do the work, not the ones who only spend their time kissing the posteriors of the beancounters, already know they have to constantly change the way they do things or else their career will slow down. Which is usually fine with IT types, as we all love to work on the new stuff, we thrive on the challenges.
* - UAT - User Attitude adjustment Tool... typically a long piece of hardwood, like a louiville slugger or cricket bat, or for particularly thick headed users... a metal bar such as a crowbar or tire iron.
Discretely Managed Virtualization was too hard to say..
so they made it Cloud Computing. You are right that it adds a layer of complexity and required expertise to handle it. Once they get third parties all over the world to run your errant virtual machines, you'll have no flippin idea where your apps and data are. Unless of course they implement a Foursquare for virtual machines. "Hey, it's running right next door.. ohp, no, now it's in Bejing, I think." But it will all be secure of course and very well managed.
I don't know about "death of" but cheap SAAS (Kashflow, Google Apps Premier, Salesforce, online banking... etc ) plus, of course, Macs, have provided this very small company and, I'm sure, lots of others, with 24/7 worldwide fully on line high speed instantly scaleable effective operation and visibility of all business functions with no IT payroll at all.
@zef re MS sponsorship
It's probably worth nothing that there is a big difference between 'sponsored by' and 'commissioned by', and the former applies here. The content of the paper is derived from formal research and the two authors (Tony Lock and me) developing and fine tuning the analysis through presentations and workshops at various IT management and architect gatherings. In order to pay the bills and fund the work of writing papers, we invite sponsorship, and we make no secret of that. The model is no different to Formula One teams receiving funding in return for sponsor logos appearing on their cars.
Hey Tony, you know that makes us the equivalent of Lewis Hamilton and Jenson Button :-)
Alright, so there is a limit how far we can take that analogy :-)
Seriously, the paper was conceived long before MS became involved, and while I would expect anyone working for a SaaS or cloud hosting pure-play to disagree with our our more balanced analysis because of their vested interest in bigging up the revolutionary approach, we have so far had no push back from people actually working in mainstream IT departments.
But let us know if there is anything you disagree with or anything obvious we have missed. The process of analysis is an ongoing one and is dependent on critical input.
Thanks for the clarification. I liked your paper as I said and do agree with you on most points.
Coldsteel German Rationality Says
1.) There is a lot of potential to rationalize the operation of Email Servers, Time reporting systems, Room booking Systems, Project management systems, standardized ERP solutions, Human Resources management systems.
A vendor like salesforce, Microsoft or Google knows much better how to operate a server system they developed *themselves*. All sorts of economy of scale come into play. SMEs don't have that.
2.) The telephone system is a raging success, despite looking very much like "the cloud". People entrust all their secrets to it, especially when they talk to their "best (female) friend".
3.) The banking system is also a system of centralized accounting we happily entrust lots of business and personal secrets.
4.) Telecoms technology has become cheap and fast. THAT is the reason for the cloud, not VMWare et al.
So it seems that the days of the SME sysadmin operating MS Exchange (or the Notes crap) will come to an end. But will it mean there is less work for flexible and educated IT people ? I don't think so.
Customizations (such as SAP customizations) must happen in the cloud as much as they happen in the local intranet. Only the current routine work seems to become redundant.
A minor? nit... maybe
"A vendor like salesforce, Microsoft or Google knows much better how to operate a server system they developed *themselves*. All sorts of economy of scale come into play. SMEs don't have that."
Microsoft? Hahahah - have you ever done business with BPOS? I quote - from this week in fact - from quite possibly their biggest BPOS client: "you guys wrote this shit, why can't you tell me what is causing the problem?" They have the same problems with BPOS other hosting/service companies have with their products, which are the same problems the in-house guys have. The biggest problem with Microsoft is that it's amateur hour - not from a technical standpoint, but from a services/business-acumen/relationship-management standpoint. IBMGS, ACS, CSC, Dell Services, HPES (you name it) would never put the people Microsoft does in front of the client - at best those guys would be technical team leads, not client-facing service delivery leads... and the boys in-house are under thumb, so there's no such thing as "not my job".
That, however, is the only disagreement - everything else I wholeheartedly agree with. The difference here is that the "cloud" solutions need to be designed/built for a multi-tenant cloud - something Google and SalesForce do, and something Microsoft does well just not for Exchange/OCS/SharePoint because those products were not designed for it... and the same applies to most of the IBM Lotus Live offerings. It's been bad enough to see these guys struggle to get their mail solutions running in true Enterprise-scale (30+K seats on a single infrastructure in my book) from their original incarnations as departmental-scale solutions. I just don't see them taking it to the next level (millions of seats on a single infrastructure) without *very* severe architectural changes.
Biggest Pile of
24/7 broadband + 24/7 uptime + proper security... does it exist? NO
WTF, more hype!
I've seen plenty of internet issues at work, including gateway server issues, and internet and site glitches can easily make servers unavailable, and I've seen how much work, thus money was lost as a result; so I can well understand why sensible IT and especially IT security people are loath to outsource servers!
you don't use these "cloud finance services", aka as "Banks" ?? I am sure it is more secure to bury your gold in the garden.
Well, maybe you have a point, considering that USGov (Roosevelt ?) confiscated the Gold of Americans in the 1930s...
Cloud finance services
Yeah, it's not like most of us are in a deep s**t lately due to those cloudy financial behaviors
Regarding small businesses, a few people have said they don't have E-Mail servers "anymore". Well, most I've seen NEVER had one, businesses that small used their ISP's email server (back in the day) and GMail or something like this now.
What is cloud computing? I've seen a few definitions 1) Basically, using virtual machines and some infrastructure so they can be created and destroyed easily, and hopefully moved from one machine to another. This was called "utility computing" back in the day -- it's not new, IBM mainframes have supported all of this since about the early 1970s. 2) Hype for whatever else the company feels like calling "cloud" -- Microsoft is pretending that hosted SQL server is "cloud", and shockingly in two ads they've started airing recently, they somehow claimed editing photos ON THEIR OWN DESKTOP was "cloud", and remote desktoping into a home computer from on the road was "cloud". yeah, I don't know what the hell Microsoft is thinking.
Other than that -- I could see some clueless PHBs getting rid of IT staff. This happens, then they have their IT stuff fall apart, and end up hiring some It staff back, or they limp along with barely-functional systems. But otherwise, going to a "cloud" will save on dealing with physical hardware failures and such, but a software infrastructure should need exactly the same care and feeding whether it's on a physical box or a virtual machine -- and, if cloud stuff is so complicated, maybe more.
Maybe just change the job title..
we run an ISP and used to have about 50 servers in our data center, with cloud computing the servers are now down to about 8. However, simply transferring the servers to the cloud will not equate to cost savings, the software services architecture has to allow for consolidation of services so fewer servers (in the cloud) are needed... so yes cloud computing is no panacea and requires people with more programming skills (to develop integration modules) and yes we still need the skills of linux and networking guys.
Murphy's Laws of Technology
1. Old technology always sticks around despite the new technology.
2. Create something that even a retard can operate and only a retard will.
3. The more complicate a specifc technology becomes, the more people will be forced to use it.
And as I like to say: it is 2010. Where is my flying car?
Not sure where you have been for the past century.
1.) The assumption that a vendor knows more about their own products is only marginally true, collectively i'd say they actually know less than their customer base. If they knew more then they would find all the bugs, and back doors to their products.
2). Once you select a supplier in the cloud, how easy do you think it will be to move.....Really....baring in mind our industries history of locking in customers whilst paying lip service to open standards.
3). Realistically how safe do you think your data will be from accidental exposure, hacking, requests from the DoJ for disclosure to the hosting organisation. A hosting provider will represent a massive target for hackers, why go after a single company when you can get 10.
4). Yes the telephone network is a cloud, and not very secure, why do you think governments use separate networks, or overlay their own security on top of the channel.
5). Banks are not clouds they are store and forward messaging services (Think about it) that overlay their own security on top of public networks, who have a vested interest in keeping data secure, and they aren't too successful at that.
6). As UK Gov. has just found out some of the IT companies are so big, they can dictate to you the terms of business, and don't really care about you as a customer, even when you spend billions with them, you still only represent a fraction of a % of their business, so once they have you why should they care about your service, baring in mind the costs of reversion, or migration to another supplier, who will by the way be offering "industry standard competitive" solutions, or lowest common denominator.
7) Clouds make sense for small business, Amazon & eBay work well. For more serious users, the bigger you are, the less sense it makes, better to operate your own cloud.
If anyone believes I'm going to trust *my* corporate usability on a UK internet connection can go fish.
The companies that run the infrastructure are pathetic in the UK and stability is NOT their watchword, it's profit.
Go fish suckers, I'll still be running when you have packetloss up to your ears.
Au contraire: Jobs for the Boys
The cloud will result in a net increase in jobs for no increased functionality - it's great!
Service wrappers will be developed for existing applications (COBOL included) and new applications will be coded to be SOA compliant even if they don't need to be. ESBs will be built and the servers to run them and the storage to persist messages. There will be loads of jobs for infrastructure architects trying to manage cloud service level characteristics and back end jobs from competing service providers and then there will be multiple instances of every cloud to cope with data retention and security requirements in different countries.
Cloud computing is a great idea because it can provide increased flexibility to the business but managing the infrastructure/non-functional side will require more and more resource while all that flexing will keep the functional guys occupied..
- Batten down the hatches, Ubuntu 14.04 LTS due in TWO DAYS
- Samsung Galaxy S5 fingerprint scanner hacked in just 4 DAYS
- Mounties always get their man: Heartbleed 'hacker', 19, CUFFED
- Did a date calculation bug just cost hard-up Co-op Bank £110m?
- Feast your PUNY eyes on highest resolution phone display EVER