Cloud in 2011: A bright new dawn... An awful lot has been written about cloud computing in recent months. Big vendors are climbing over each other to claim an understanding of cloud and 2011 is the year it is supposed to go mainstream. Cloud computing will save us money, it will simplify our IT systems, it will transform the way we interact with government, it …
I was asked if I could update a tiny website I look after for a youth group, bring it up to date and all that stuff. Looked around at similar sites and found that most looked better, more 'modern' but that was where many had stopped -- 2 or 3 years since the content was changed rather than the pretty side of things, empty pages - the standard half-built, rarely touched CMS sort of thing. Ours - functionally it's far better than most but doesn't have bits of Script flying around and looking pretty.
Some bright spark mentioned "Cloud - put our photos in the cloud ".
"What, you mean put the main photos on a remote server and provide links and thumbnails locally?"
"Yes, you've got it"
"But they've been on Flickr all the time with links and thumbnails on the local site, it even says that in the text."
"Yeah, but that's not 'Cloud' is it?"
"Oh, what is?"
silence
10 years ago I was pushing the IT department that I worked in to move to a utility computing model (aka the "mainframe" model 8-) where the IT department takes the risk on capital costs and sells compute cycles and storage blocks and scales upwards/downwards and outwards/inwards as necessary .. but nope ... "customer buys a server" rather than "customer buys a service" prevailed and I blame it on Microsoft and the PC on everyone's desk solution and dodgy software stacks. Earlier, the physical cost of X Terminals killed their own market when 'low-cost' PCs out banged them per buck; shame as PCs also made reasonable X Terminals once you cut-down windows and auto booted into an X Server.
Nowadays its coming full circle again with VDI and cloud being the new mainframe, albeit with worse security but lower licencing and better terminals 8-)
...and from a L2/L3 headcount standpoint it's going to wind up being a wash. Babysitting is expensive :P
One thing that's not called out in this article, is the loss of control when you move from on-prem to a Cloud provider is a huge change. SLOs be damned, when we had an issue in the old world we would have it identified and a clear idea of the impact in very short order - on average I'd say within the first 30 minutes - and our MTTR for critical issues was somewhere on the order of 1:15-ish. There was one throat to choke for leadership - I owned it.
Nowadays, for major issues we're lucky to get a response back in the first hour just to say "we're working on it" and we're lucky to ever get a proper impact statement (typically coming 7 days after the outage in the form of a PIR) and usually go through entire outages guessing about how many users are affected (i.e. what the hell is going on)... and don't get me going on lack of proper after-action on their RCAs.
I think the assumption usually is that the cloud will be more reliable and trustworthy than on-prem (after all, who is more qualified to manage Exchange than Microsoft... right?) but that's not *always* a good assumption in my experience.
.. to recognize another 'bubble' forming when I see one. Sooner or later they always go 'pop' leaving a lot of people crying into their hankies. Main issue is that broadband lines are nowhere near reliable enough to place so high a dependence on them.
MIght add that if a site needs Messagelabs filtering then they need to look-at WHY they're getting all that spam. Spam doesn't arise out of nowhere, it comes from the firm's addressbook being harvested through one of several very common security failures. Fixing the underlying security problem will be cheaper than filtering, and will lose less valid mail.
Security failures leading to address books being stolen? Dictionary attacks (the quickest way to "steal" the address book by trying a lot of address and recording the NDRs) are going out of fashion :)
These days with the power of your botnet (a very successful application of cloud computing I may add) you can miss out that step.. The spammer simply need to start with aaaron@ and finish with zzacharia@ and keep trying to deliver the mail.
Mind you I don't disagree about the bubble bit.
“I see the cloud as the start of the end for IT professionals. Once businesses start putting their systems onto the cloud there's going to be next to no support needed"
I don't think so .. I recall the same kind of thinking when people migrated away from the mainframe to the more ease-of-use graphical Windows platform. You'll still have to hire people to maintain your virtual servers in 'the cloud' ..
--
Where did my billg smiley disappear to?
... happened years ago?
The article touches on what maybe a key turning point: do consumers (possibly not too IT literate yet server experienced say work related experience) want it, need it, find it attractive?
There will never be a demise of the IT Pro - just a shift in duties and responsibilities.
IS part of the core business. An accountancy company that eliminated IT spending would have to spend just as much (almost certainly more) on staff, ledgers, notebooks, pens, pencils and ink. To say that IT isn't the core business is either a deliberate attempt to push the Cloud case, or a fundamental understand of exactly why IT is used in a business in the first place.
The ONLY reason to make a change in how IT is used by a company is to increase security, increase functionality or decrease costs, and none of these should be done at the expense of the other two.
Cloud (really just hosted services) might decrease costs due to scale, but does it do so without impacting the other two? We have already seen Cloud outages that have severely impacted businesses. As more people move to the platform, the more critical an outage will be, to the point that a single outage will have a measurable impact on the GDP.
And don't argue that companies have outages with their in-house IT. A Cloud outage eliminates all services at once, for multiple companies. Few companies hosting their own servers will have ever suffered a complete outage. One application might go down, or one network segment. But never the entire thing at once, short of a power outage (which would take the desktops with it as well) and that usually stops all companies in their tracks.
Cloud is a way to put all your eggs in someone elses very flimsy, untested basket in order to save a few quid. And that saving is eliminated with the first outage, because the stress and downtime will cost many times over.
Using gmail is using the cloud. Huh this sums up how people have been trying to sell so called cloud services, at least those that I have heard from.
To me the above is just server client. Yes a very bit fault tolerant server but essentially just client server. Running your services in this so called cloud is once again client server, but your servers are rented and probably more fault tolerant than what you can do yourself and scaleable.
There are obvious benefits to the model but it is still client server. Or is no one connecting to this imaginary cloud.
Once you've given away the keys, can you really complain once your data is hacked and mined? At least when you host your data locally you have full control over access and control. Once you give that away, the best you can hope for is than the cloud solution provider you choose will pay up on massive data breaches. Though they probably won't because they'll have that covered in their EULAs
This post has been deleted by its author
Really will be the death of the IT department as we understand it. They say it will be innovative? Anyone who has outsourced knows that innovaton is stifled in a wrestling match between the services delivered, the services required and the cost of implementation and training as you have to ensure that the host organisation is up to speed to support the client.
Local IT has shifted many times before, it will shift again into a new and un-inspiring role of becoming sales rep's and service contract managers. Yes we would all need a degree of local support but the reality is when your network consists of terminals, switches, routers and printers, local support is going to look like a networks guy. The main relationship between host and client will be managed by a service manager on both sides, and the rest of you can get stuffed or enjoy your new roles in accountancy.
As a local organisation IT guy... I mean sales rep... You will enjoy life as a person who's task is to no longer develop an application or database hack or even patch a server. Your job will now be, how much will this service cost and in order to deliver on that you're going to have to talk to the Relationship manager. Hell why not cut out the local guy too, uh, I mean streamline the operation...
As far as new innovations are concerned, explain to me how this wil work when you are not only no longer in possession of local support to implement or even develop such ideas, when you no longer host your own data and no longer even host your own applications. Will you invite the host organisation's developpers to your office at great expense to ask them please mr company develop X for me so that I can be more efficient but make it so it doesn't cost me an extra $100K on my contract? Innovation will come in the form of contract re-nogotiation as the host and the client wrangle over costs of developing and implementing any solution they need to deliver the new service and hope to christ that sufficient infrastructure investment has happened to deliver on the requirements.
Cloud computing will create 2.4 million jobs? I don't doubt it, the reality of that statement is that those jobs won't be entirely new and some of them will be making many more jobs redundant. When your only tool is a hammer, pretty soon everything starts looking like a nail. The axe man cometh and I wish you all luck in your new accountancy positions.
The primary problem with cloud computing is that the edges of the concept are fantastically fuzzy, and so is the definition of containment. You need access to data to process it, but as soon as you start distributing that you hit containment and control issues. The part that IT directors and other wannabe *cough* innovators skip when they buy the latest buzzword is that they have legal obligations to the company and the law in general to protect information from theft and disclosure.
Do you really want the latest airframe drawings in a place you don't know who has access to it? Do you realise that exporting personal data across borders is in some countries simply straightforward illegal? As an example, do you really want bank data spread across the planet? If not, are we now talking about cloud slices?
It takes intelligence and skill to decide precisely what cloud computing can be used for, and where it is inappropriate. That skill resides in professionals. I can see the profession take a hit - and then present the bill when it is time to clean up the mess. There won't be enough staff to fight the fires when cloud computing turns to clod computing.
Oh, you want evidence? Look at the major disclosure problems at the moment - always large networks. And they're not even fluffy..
<rant> And cloud allows multiple ancient highly customised mission critical apps running on old and odd backend databases or hardware to run how? Oh, only M$ or basic *nix stuff. Yeah right, really useful for big enterprises.
Another idiot buzzword salesweasel campaign for chronically gullible PHBs. Great for sales weasels though. Much to promise, since they never have to deliver. Lots of potential for bean counters to sack staff, since bean counters are never crucified for stuffing up an organisation. Bigger performance bonuses for the board and senior suits because their pay has no connection to a company surviving more than two years after their looting. One could argue that incompetence is mandatory at that level, judging by the last 3 years of examples.
The loss of control also means loss of responsibility. Perfect for the cowards, weaklings and process droids that masquerade as business people.</rant>
When I previously worked in telecoms I learned that there are many single points of failure that are often overlooked. One of those is your relation with your supplier. When you are in a high-volume business there are always going to be disputes with your suppliers. At that time you do not want to be at their mercy where they can shut down your whole operation over a payment dispute.
So basically you need to have redundancy of cloud suppliers if you move your critical operations into the cloud. Once you have that redundancy it becomes less important if one of the cloud suppliers fails, so you might as well move back to renting co-location space from different suppliers in different locations using different internet backbones. At least that way you can make your own risk analysis / mitigation plan and bargain with your suppliers on price/quality.
Another thing, I didn't notice TheRegister covering this, but if Amazon - the mother of the cloud - cannot allocate enough resources to sell a single MP3 from their new music locker service that they are trying to promote... why should we trust them with our business critical services?
http://www.computerworld.com/s/article/9217020/Lady_Gaga_s_Born_This_Way_overwhelms_Amazon
"So basically you need to have redundancy of cloud suppliers if you move your critical operations into the cloud. Once you have that redundancy it becomes less important if one of the cloud suppliers fails, so you might as well move back to renting co-location space from different suppliers in different locations using different internet backbones. At least that way you can make your own risk analysis / mitigation plan and bargain with your suppliers on price/quality."
By the time the brown stuff contacts the air-circulation unit the cloud competitors you were thinking of using have been taken over by the first lot who have 'centralised resources' and what seemed to be an option is actually running on the same farm with a different label on it.
First comes the innovation, then comes entrepeneurs(sp?) and the fast buck crowd who don't give a toss and only see money.
Shut down smaller server farms, relocate to a more 'understanding' economy (for a few months) then do a runner. It's the next 'property ' boom, if they can get away with it.
...it is a good time to define what this "cloud"-ed environment really is?
The way I understand it, services are outsourced to a partner who in turn outsources it to other providers, and in the end the original user has no idea where his/her data and computing power comes from, but only sees the results...
Is this the perception of others as well? Or are we all confusing the current "cloud" buzzword with (flexible) space/power rented in a data center?
(WTF because I am really getting confused... ;) )
This whole thing about the accreted widespread use of the word "Cloud" and what "cloudiness" implies is unfortunate. Why not say "Remoted", "Offsite", or "Automated Storage Servers"?
Unless the mystique and mystery of cloudiness is somehow an attribute, and I can't see how that is useful, get rid of the use of "Cloud"....no wonder IT professionals are unfurling their black umbrellas.
Yes, there may be small companies and privately owned manufacturers of a single widget for whom a public cloud solution might be beneficial. In a complex enterprise, the reality is very different.
There are many short-term, accountancy and bonus-driven CEOs and FDs that will look at the bottom line without understanding the loss of flexibility in preventing their enterprises from integrating multiple, disparate data sources and systems, and killing off any opportunity to independently manipulate and manage business processes without having to go cap in hand to a cloud-based ERP supplier. For most large organisations, what the cloud delivers in terms of capital efficiency, it more than takes away in loss of response and flexibility to changing business needs.
For the most part, the companies that go to public cloud will do so with little in the way of strategic, architectural governance over IT. In the long term, that may well cost them a competitive advantage.
See the article: 'Not Every Cloud has a Silver Lining' at http://sitfo.wordpress.com
One thing that often seems to panic big businesses and such is getting hacked by the Chinese, Iranians, Korea or similar for industrial espionage purposes, but if they outsource all the data storage to the "cloud", can they really be sure where the data is actually being stored, and thus who can actually read it?
For example say salesperson X comes from company Y selling this cloud concept and you'll have data stored in our cloud which is replicated in several geographically distinct locations for extra safety. If they forget to ask where these extra locations are, what's to stop the government of the country hosting the servers peaking into them and supplying the data to companies in their own country?
Amazon might be open about where they are, but not all will be especially if say the Chinese gov. starts offering discounted rates on bandwidth, data centre space and not worrying too much about all that environmental stuff.
I agree with Anonymous Cowards who would say that Cloud does not replace anything existing, it is just an additional, extremely sophisticated novel application which just sits on top of, and takes over Command and Control of Intelligence and Secret Intelligence Services, and with traditional government models evolved to do its bidding, with the pumping and pimping of its IT and Media Rich Programs, does Cloud Hosting Advanced Operating Systems deliver a Future Friendly Virtual Order ..... in New Virtual World Order Regime Programmings.
More on such as inevitable, because it is actually a present program running "black and hot" in most definite, Alternate Reality Circles/Round Table Organisations/Serious Gaming Teams, can be found posted in a comment on 05/30/11 04:11 AM here ...... http://thedailybell.com/2417/Queen-Worried-About-Empires-Meltdown.html
Although one might suppose that such as will be, is never ever a surprise or unnecessary troubling worry to HM and Co., even should some upstart pretenders, bolstered in their ignorance and arrogance with imaginary stores which are hoards of artificial flash cash/new money, might presume that they can make it so.
"Cloud Computing" seems to an oxymoronic addition to our already overcrowded vocabulary zoo which we delightedly call "computerese". In a scientific attempt to be precise and have boundaries of responsibilities, and attempting to set standards/methods of....computing!....costs, Cloudiness seems to prevail.
I KNEW this was coming, because in the late 1970's I was in a company just then becoming 'computerized', and the haughty Programmers all but demanded that we kowtow and flagellate ourselves, requiring that we the unwashed submit to their way of doing things....I.E. their newfangled humming, hot machines were not to serve us, but we were to serve their Oz-like masters, the.....(cue fanfare!)...Computer.
Hence, at age 79, I can sit here at my laptop and click out my triumphant declaration that Th' Virtual World now seems to be hoist on its own petard.
So, just who/what is in Control?
HAL is stirring, and beginning to flicker, and smirk.
"You could end having your commercially sensitive data in hands to foreign coppers who may/may not sell the data onto their business pals." .... Ray 8 Posted Monday 30th May 2011 23:26 GMT
Yes, that is possible, Ray 8, and also quite probably most likely, but the smarter foreign agent will realise that there is the very real probability that said commercially sensitive data has hidden and invisible back doors for trojan horse action which relieves business pals of their control of data, for some juicy morsels found, with or without a warranted search, are deliberately left for discovery as delicious bait .
"“In general, small firms in the UK are smaller than those in France and Germany, and therefore less efficient and less able to make economies of scale work for them,” says Hogan."
Definitely some fuzzy wuzzy logic being used there, methinks, as no matter how small, or how big the firm may be, does everything depend upon the best brain delivering new IP to driver it. Although the best brains may choose to work for the best biggest companies to develop their IP with third party resources...... and get paid a fortune for it too, which is a very attractive, and even perversely addictive bonus to many creative driver types.
* .... that was before future intelligence became readily available. Now is that old adage and situation much more the reverse, with those who would wish to know what they need to know, chasing that which needs to be known via engagement with those who know it ........ :-) which is essentially exactly the same in reverse too. Hmmm :-) How beautifully weird.
> A very large number of small firms in the UK are so small that even a server is an unnecessary expense.
A single instance virtual server would be useful for the smaller firm, if only for automated backup. Speaking from experience, you're dealing with people who can't remember to , once a day, swap a tape cartridge.
If you're running your business or connecting with your customers through the cloud, then these recent outages aren't a very good impression. They should rename it the Almost-Always-On-Cloud ..
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
