Servers

Note to CIOs: Get your head out of the cloud

Patrick Gray takes issue with the cloud-computing hype machine and extends a word of caution to CIOs who are in danger of becoming hypnotized by IT buzzwords.

"Cloud computing will revolutionize IT!" shouts an IT rag in an article laced with superlatives that, like the promise of "Free Prize Inside," a child's favorite cereal finds you digging for more only to be left with a stomach ache and a worthless trinket. Despite almost daily press releases from the vendor du jour claiming how their product or service is "cloud ready," few can concisely articulate what exactly "the cloud" is, and how it is conceptually or pragmatically different from your run of the mill, but decidedly unsexy, hosted enterprise application or service.

While talk of clouds may get marketers hearts palpitating, and make for inspiring ad copy, most CIOs would do well not to believe the hype. If talk of cloud computing feels like déjà vu all over again, it is likely due to the fact that cloud computing is, in one way or another the same thing we have been doing for decades: connecting disparate applications and services around the world through a hodgepodge of networks, often not owning or having direct control over many of the components. It has been called SaaS, hosted applications, e-commerce and other decidedly gauche monikers before, but the concepts and execution are fundamentally the same. Arguably that boring, but ubiquitous relic from the 1980's, EDI could be given a shiny new polyester suit and silver "cloud" badge with a minimum of marketing polishing.

If hosted apps did not make sense to your organization a few years ago, putting another layer of buzzword frosting on a stale cake will likely make little practical difference, and basing your IT department's strategy on "cloud computing" with little understanding of the business risks and benefits is as solid a footing as a cloud itself. You'll be well ahead of the curve, and far more respected in the boardroom if you quickly move beyond the hype cycle with the cloud, and look at how moving applications outside the firewall might have a concrete business benefit that outweighs the obvious risks.

If you want to fast-track your cloud strategy, dust off the months of work you likely already spent looking at hosted applications like Salesforce.com or Google in the past.

Buzzwords and trends have certainly sold large and expensive IT forays in the past; the executives (or even whole organizations) who chased these mercurial entities with no strategy or tactical plan that acknowledged the potential benefits and quantified the potential risks usually suffered greatly or ceased to exist. Your peers in the C-suite will quickly see cloud computing for what it is: technical and functional outsourcing. If the crux of your strategy around dealing with the cloud is a glorified version of: "But the IT press says clouds are ‘in,'" you will quickly be elbowed out of the boardroom.

Whatever you want to call it, cloud computing certainly has its merits, but behind all the hype there are very real risks to go along with the potential rewards. There were continuity, security, extensibility, and usability pros and cons back before "the cloud" arrived on the scene, and they are still just as real today, despite the new terminology.

About

Patrick Gray works for a global Fortune 500 consulting and IT services company and is the author of Breakthrough IT: Supercharging Organizational Value through Technology as well as the companion e-book The Breakthrough CIO's Companion. He has spent ...

32 comments
RealGem
RealGem

Thankfully, I am immune to hype. I've been doing this too long... But, I depend IT execs at other companies to get swept up in the hype. How else can ideas and technologies be proven. I absolutely want someone on the bleeding edge other than me, so that I can reap the benefits of stable technology later.

Brevoort29
Brevoort29

There is a potential risk with anything you do.

LocoLobo
LocoLobo

To me, it seems like a lot of people have been "Clouding the issues". In the end, when you look at the specifics of their proposals it looks like outsourcing. That is always an option. But weigh its costs/benefits carefully. Just my opinion.

smatteson
smatteson

What a nice change from the smug business columnists loudly hawking cloud technology and predicting the demise of the IT industry. "Get your data into the cloud so you can quit paying those overpriced self-absorbed techies so much money!" seems to be the underlying theme of their marketing schtick. Never mind when the company connection to the internet goes down, leaving you with no access to your information. Never mind when the cloud hosting agency connection to the internet goes down, leaving nobody with access to their information. Never mind when the cloud hosting organization has insiders combing through confidential data and perhaps selling it to competitors or blackmailing you with it. Never mind when the backups of your data aren't running correctly and nobody at the cloud organization bothers to tell you because, well, that might be bad for business. Never mind when the cloud hosting organization folds and takes your information with it, leaving it tied up in litigation. Never mind when there are problems with data, you call up for support and you wind up speaking to a different person every time, who reads from a script for six bucks an hour, and who doesn't know the company, its people or its data. The "overpriced self-absorbed IT guy" who used to sit down the hall could have had your issues addressed and resolved hours ago as meanwhile you're telling your tale of woe for the ninth time and identifying yourself ninety ways to Sunday, over and over to prove you are who you say you are. Now who's losing productivity and precious business time? Like all the other pro-outsourcing trends, some smaller companies with less capital to invest in data centers and IT staff may see some benefit from the concept of cloud computing - and suffer through the drawbacks and disadvantages. If they can't scrape up a few hundred thousand for servers and an IT guy so they can sell apples from a streetcart, then combining resources with other one-shop Johnnies makes sense. But the bulk of the organizations who understand IT is an asset and a means to grow their business (and not some tiresome cost-center draining their resources and productivity, as some business columnists pointedly insist despite reality) will keep their data, applications and people in house for the best performance, results and efficiency.

Kris.J
Kris.J

The whole 'cloud craze' lately is just crap - new marketing crap layered on old hat stuff.

gometrics
gometrics

I agree with author on all fronts. The proper question is how much control does the company want to have?

dg.itpro
dg.itpro

A glimpse of what could potentially happen with cloud computing? One article said pretty much, if something like can happen to tech giants like T-Mobile and Microsoft, what can your business expect from SaaS?

steven.mauss
steven.mauss

Sorry for the over-use of a metaphor, but I've been fighting "low-pressure" systems for some time. Our company deals with mostly very large enterprises and they are extremely reluctant to turn over critical systems to a "cloud" of any type. More likely, they will move to a "low clouds and fog" system where it stays within their firewall. Sound like the mainframe days of old? It's pretty close, but there is one primary difference: A common user interface via the browser.

CWSpence_Intel_IT
CWSpence_Intel_IT

Good message to get beyond the hype and really understand the business reasons & risks of using Cloud. Companies really need to think about which applications are best served by outsourcing entirely or run in the Cloud. For SaaS, these are large scale, commoditized applications with industry standard workflows. No intellectual property or sensitive data.

richard.wilson
richard.wilson

Finally, someone says what I've been saying all along: "Cloud computing" is nothing more than a fancy word for outsourcing.

jmarkovic32
jmarkovic32

It goes a little deeper than buzzwords, IMHO. CIO's believe they have to be the CFO Jr. or CFO Lite in order to fit in in the board room. Some of them can't help it because they are nothing but ex-bean counters in the first place. Everything they do is about cutting IT costs instead of IT generating revenue directly or indirectly via process improvements. That's why we need more C-level techies who hang up the engineer hat and step up into a strategic role.

drmorton
drmorton

Finally someone who knows what is going on puts a dose of reality back into these "stupid" ideas. L P Morton, Ph.D.

Deadly Ernest
Deadly Ernest

but been very specific about what the problems are. I will also say the biggest problem is EXTERNAL cloud, not INTERNAL cloud. An internal cloud computing environment, which will be little more than the old style mainframe or master / slave or Citrix type set up but use a server and local desktops using browser to access the internal servers would be a good idea for many enterprises. However, no CEO or intelligent person will take their enterprise critical system to an external cloud and make themselves dependent upon the operation of the whole internet and some company or companies that can be anywhere in the world. If they do, I know some people with some lovely waterfront land to sell in Florida, I can get a good commission on the sale. Many security issues would affect an external cloud set up, let alone acces, and reliability.

sammyle
sammyle

I agree 100%!! I have had to deal with justifying why we shouldn't move to a "Cloud" all year. Frankly speaking, the typical IT candidates who are promoting Cloud are all over priced and just simply trying to pay off their own Capex investment on infrastructure. Any small to medium business would be much wiser to move to a Virtual platform in house (lower annual reoccurring costs compared to any cloud solution) and over the next 3 years recruit the right applications for their business that are workable in the cloud. By then the big guys would be past the 3 year Capex pain and the economies of scale should start to tip in our favour.

b4real
b4real

You bring good points - but for the real decision-makers, you and I included, would we not enter with the requisite caution? Cloud solutions are going to be successful, whether that be a niche or more. Fact is, "cloud" is old-school in a way. A PSTN is a cloud of sorts, digitalriver is a cloud, there plenty of examples so far.

philr
philr

I visited an HP facility in Santa Clara many years ago. They had three networks. One for admin stuff, one for the work group and one for external communications. Each person had access to these via different hardware. They understood how to mitigate risks. I cannot see them ever using an external cloud. FWIW the notice boards were really interesting to someone like myself. The person showing me around did not know I had been in IT. :-)

slikone27
slikone27

I was reading some of the comments on this article and this question just popped in my head. How old are you guys? Times are changing and Cloud computing is not what it was years ago. With HTML5 on the horizon web apps will be very comparable to desktop apps. While it is not here yet and you will need some type of mix of both the Cloud and desktop apps this is clearly the future. I guess there are always people that will fall behind the times. There are always people that need to have their spreadsheets because they don't understand databases or simply don't want to use them because they are comfortable with what they know. Ask yourself this (especially small and medium sized businesses): Can you really provide the reliable and redundancy that a company like Google has? How many data centers do you have? How are you backing up your data? If the focus of your business is not IT then is it worth spending a lot of money every year trying to build your infrastructure? Do you have a dedicated security team monitoring your networks 24/7 watching for intrusion? The answer to at least a few of these are probably no. I am not trying to blast anyone but things are more advanced than they were years ago. Technology is changing and this is the path of a new generation. Now, if your company can spend hundreds of millions of dollars a year to setup your own private Cloud then you might want to keep everything in house. If it can't then you may want to take a look at what Salesforce, Google or any of the other players have to offer. Even Microsoft is trying to get into the game. Bottom line is their infrastructure is probably more secure than yours whether you would like to admit it or not.

CG IT
CG IT

They postponed moving the entire system to a cloud provider but now they are in the process of moving email to Google including the police force email. Cost of the contract for hosted email from Google is about $7.2 million. City of Los Angeles believes they will save more than that using the hosted solution. Note: the City of Los Angeles doesn't use Exchange for their email system. What Google hasn't said was if the servers and infrastruce will be like the DoD model which is seperate hardware. More likely they will be on the public cluster.

steve.rayner
steve.rayner

When I began in the ICT profession 35 years ago it was in a bureau and we ran systems for customers. Several years later the same function was "hosting". Then came Application Service Providers; when that catchy name didnt draw hoards of companies running lemming-like with open cheque books it seemed to simply fade away. Now "the Cloud" - its bureau services by another name. Years ago we were responsible and accountable to our clients who could come and see what we were doing (and audited us regularly). Now we seem to have an amorphous mass into which I am led to believe I can pour all of my operational cares and woes and magically all will be fine. Now I may be a cynical old man (and a CIO) but we have at least four names for essentially the same thing and I am always suspicious when the name of anything is changed so I ask "what is behind the hype?".

oberois
oberois

After reading Patrick Gray?s article and several posts I believe the name ?Cloud Computing? be changed to something more befitting ? ?SMOG COMPUTING? (Superbly Magnified and Over-blown Grand-Scheme Computing)

viveka
viveka

The net business effect of a cloud is cloning the same solution. What this will result is a lot of undocumented ghost min-applications where a lot of IP is stored off system (but may be on storage systems). This occurs, because in reality, one size does not fit all. Also, another off-shoot would be employees using free "open-systems" to do their work as the enterprise will no longer support these needs. It would be fun to watch the new set of challenges this will create.

Deadly Ernest
Deadly Ernest

for the latest variant of the cloud is they push the need to use everything off their servers and pay them for it. This moves ALL your data and operations outside your borders, outside your control, and outside your security. That didn't work for Czechoslovakia or Poland between 1945 and 1991 - it put them at the total mercy of the USSR, and going external cloud puts you at the mercy of the company you go with and the Internet, and the idiot down the road with a bob cat digging holes in his garden.

Deadly Ernest
Deadly Ernest

and safety inherent in the basic design of it. What Google Apps provide is just another version of thin client, but done on a browser. An internal cloud need only be a case of running the browser based apps on a company server inside your network, just the same as you currently run a thin client server. There is a lot you can do on the web, and there is a lot of screws ups currently done on the web, but the web does NOT provide security or safety, it was never designed to do so, and it can't. BTW I'm in my 50's and do prefer to use the best available technology for what works, provided it's going to be around and doesn't breach security - I spent too long working in security organisations and environments. Before you suggest people go to web based operations with the data stored out there somewhere, you should read some of the many things that can cut you off from your apps and data, and how many different ways they can lose it all for you, or even sell it to your competitors for you.

Deadly Ernest
Deadly Ernest

putting the email outside at Google, is going to be one hell of a security issue as they can't guarantee the security of the data or restrict the access once that's done. And it gets worse with other data.

CG IT
CG IT

ROFL LMAO :) Well Said!

CG IT
CG IT

email system. That's 33,000 employees including the police department. Police will be the last department migrated over. The contract isn't signed yet, there is a clause they have to agree on. That is monetary compensation to the City of Los Angeles should there be a breach in security...which means Google will shoulder the financial risk if the City of Los Angeles is sued . Google will have dedicated servers to the City of Los Angeles for the email system. Can't see them on the general gmail system. No outages for that.

Deadly Ernest
Deadly Ernest

see a huge pile of legal and performance, and later cost, issues with an external cloud.

CG IT
CG IT

that pull all the strings. That network is, to say the least, so full of specialized programs created by the "horde" that no one really knows what's what. You've got the "horde" that are indespensible because they are the only ones who know how it works and how to fix it. It's to costly to change over to something that is standardized so Cloud looks really good for them. Saves lots of costs and gets rid of the specialized "horde".

davidmastro
davidmastro

I could listen to train wreck stories like that all day... there's some good info you put up there that I may need quote you in the future when I need to convince others to keep there eyes open. Thanks for taking the time.

Deadly Ernest
Deadly Ernest

good cloud computing is, I think of that very old rhyme. For want of a nail a horse shoe was lost, for want of a horse shoe a horse was lost, for want of a horse a rider was lost, for want of a rider a message was lost, for want of the message a battle was lost, for want of the battle a war was lost, for want of the war, a kingdom was lost. Thus, it is clear the king lost his kingdom because he didn't take time to check all his men's horses were properly shod before heading off into battle. The dominoes will fall regardless of what you want or think.

Osiyo53
Osiyo53

I used to work for a major telecom. Saw numerous such incidents and heard about more. Chuckle, in one incident state contracted workers were involved. Digging along a highway they severed a major trunk (fiber optic) not once, but 3 times in a single week. Each time they isolated a sizable area, numerous towns, from the rest of the word. In another incident, a private contractor working down in Texas severed a major backbone segment that put down ans isolated an estimated 10,000 businesses. From Texas up and along the East Coast. Of course, in each such incident we were able to make a repair in a matter of hours (from 6 to 12). But during that time a lot of customers had their activities severely curtailed or coming to a screeching halt. For wireless users and advocates who're not tech savvy, I'll note here that your wireless connection only goes from you to a nearby radio tower, and from there your data goes over either copper or fiber towards that far connection. So cut copper or fiber lines affects you, too. And, of course, outages occur for other reasons also besides the guy using a bobcat incautiously. There are earthquakes, floods, storms, etc. i.e. Back several years we had a major flooding incident out in the western portions of the state I live in. A lot of stuff, including whole towns and portions of sizable cities got submerged. Within the affected area there were towns and such that were okay, but ended up isolated. There were roads that'd get through to them, but both voice and data comm was isolated to only their own small area. Communications with the outside world was pretty much restricted to emergency stuff that was routed through some volunteer amateur radio operators. In another incident, a major ice storm downed over 10,000 telephone poles, blew power transformers, etc over a wide area. Hundreds of square miles. It took a while for us to get voice and data comm restored. And then there are simply those "Oops" incidents. Like the one where where an independent contractor was installing some new equipment within one of our major routing centers. They made a bit of a wiring error in the power wiring for the device they were installing. And "Blam", we had a nice fireball and arcs and sparks. And that routing center played Dead Cockroach from that early evening til about 10:00 a.m. the next morning. Their error toasted parts of a major electrical distribution panel in the building and blew the transformer outside. Of course, emergency crews responded immediately. But repairs of such a nature are not accomplished fast, throwing more people at it doesn't get it done any sooner, they just get in each others' way. And needed replacement parts are NOT something one keeps on hand at every place they might be needed. We had em, but at central storage/supply locations. Some were both very pricey, and the kind of things yah rarely need so its not as if you buy 1000 spares and keep the spares at every possible location they might be needed. Some came from a location 150 miles away, a few had to be flown in from even further using a chartered aircraft. This outage didn't just affect our lines and services, other companies that were in the voice and data comm business relied upon equipment in that building for routing services. Fortunately the area was largely rural and semi-rural. Contained numerous towns and a few small cities. But within that served area were located such folks as a major IBM data and support center, a world-wide recognized major medical center which provides among other things around the clock consulting and advice to doctors world-wide, etc. So the effects were felt far beyond the boundaries of the outage. Chuckle, we had CEO's (or their reps); city, county, and state officials; reps from a couple of Federal offices; the FAA, etc, etc, etc calling us or making personal visits wanting updates on the status of restoration of services, explanations, and so forth. It was kinda humorous actually. In this case I was a major player, one might even say the key one. It was one of my facilities, as concerns maintenance and repair. I was the one who knew how everything was connected, what was what, what was required, and so forth. And at one point I headed for my work vehicle announcing to anyone interested that I'd be back later. I got stopped by several people for a moment as they asked just where the heck I thought I was going. I shrugged and commented that I was going for some coffee and some breakfast. I'd worked a regular day the previous day (it was oh-dark-thirty about this time), and had just gotten home when I got this emergency call. Had now been on this site for about 8 hours. I was tired, and hungry, and my ears hurt from all the constant prattle, demand making, complaints, etc. Right at this moment, the workers didn't need me as I'd given em enough directions to keep em busy for the next couple hours without the need for any more input from me. I'd been rotating work among the workers so that they'd had some breaks. Now, it was my turn. Chuckle, I had someone ... don't remember the name or exact position held (and was too tired to care) but he got real irate and demanded that I halt and stick around until everything was back on line. Otherwise, in his opinion I was derelict, incompetent, and a bunch of other stuff. And he growled something about seeing to it that I was fired if I left. I just stopped and looked at him and told him that if he thought he could do a better job than I was doing ... be my guest, have at it. Fine with me, meant more time I could take to have coffee and breakfast, could maybe even sneak in a nap. So, "Kiss my a**. Do what yah think yah gotta do, but I'm having coffee and breakfast." LOL ... the guy looked like he was gonna have a heart attack or blow up literally. I ignored him. Drove down the road a bit to this truck stop I knew where they knew the real meaning of "Strong coffee, please.", and which served some tasty eggs and hash brown potatoes ... and plenty of it. I noticed that some car from the site followed me and a guy from it followed me inside. Turned out to be some rep from IBM who'd came to the site to get updated on the problem and status of repairs. He was something in upper management, but turns out that he used to be a hardware guy back when. He just smiled and introduced himself, then explained that he'd been in my situation before, and knew what it was like. He said he was buying the coffee and breakfast, and having some with me. "Forget that idiot back there, he's clueless. Probably just a pencil pusher who has never done anything but that." Turned out to be a nice, reasonable fellow. This outage was undoubtedly costing IBM and great deal. But he understood that "SH*T happens", and the best you can do is deal with it. And maybe come up with some way of avoiding the same in the future. In this case, the telecom I worked for approached all parties concerned afterwards to (1) get permission and (2) seek some financial assistance to install a mirror site some miles away that'd automatically assume duties of the first site upon a failure and vice-versa. What some folks don't seem to realize is that while all these networks have alternative routing, multiple trunks/backbones, and so forth. Any particular location and the business or person served might be at an end point that has only one path out somewhere. A bottle neck. And if something goes down at the bottle neck, you're screwed. Your data or services might exist on multiple sites at multiple locations (mirrored) but you're not gonna be able to connect to any of them. I now of one business that learned this the hard way. Now they have two separate land line connections, both fiber, PLUS on top of their building they've got a satellite link just in case. Satellite link is slow, but at least works. A nationwide firm, they have the same setup at every location. Plus their data is mirrored locally on servers so that if all else fails, a site can continue working and once a link is remade software automatically synchronizes and updates local and wide area data servers. Might sound like an excessive move on their part but they analyzed the cost to them of outages in terms of both real money lost plus reputation lost with customers and decided the expense was worthwhile.

Deadly Ernest
Deadly Ernest

city in Queensland a few years back, they guy didn't check for cables and ripped up the main trunk line connecting the city to the rest of the world, no outside comms for two days, until a whole section of cable was replaced and connected up properly.

Editor's Picks