Servers

Cloud solutions: Will they be worth the hype?

While there are few specific cloud solutions available, directors and managers are setting their sights on what will work in the cloud. In this blog, IT sage Rick Vanover spreads the field for gauging what this can mean to underlying technologies.

It is pretty clear that TechRepublic members are divided on the cloud, evident by my posts on Gmail being attractive and my opinion that the cloud will prevail. So, now I’m going to play the critic. I’m convinced solid cloud solutions can be successful only where there are clear, technically defined solutions and a cost benefit. Further, I feel that only "big" players will be legitimate contenders for the larger organization’s business.

Beyond the Google Apps messaging service, I’ve been poking around some of the other players in the space. I am a details person, so I find myself looking at individual technologies to be components of cloud computing solutions. Google has Web technologies down, and I think the Google Apps messaging service will be successful. Including Google, I feel the cloud potential is great. Some other technologies include virtualization, robust file systems, flexible security models, and presentation technologies. To get a good picture of a comprehensive infrastructure of what cloud computing looks like, check out this Sun white paper on cloud computing. This document does a good job of identifying the underlying technologies, but it doesn’t give a specific cloud solution. This is somewhat contrary, as many cloud providers at this point are a little light on their details.

Part of the division of TechRepublic members is the fact that we are for the most part working with hype only. Once we have specific solutions on the table, are they going to be worth it? I believe in the concept of cloud computing, but we have to have a good handle on the underlying technologies. Share your comments below on where you see the cloud going.

About

Rick Vanover is a software strategy specialist for Veeam Software, based in Columbus, Ohio. Rick has years of IT experience and focuses on virtualization, Windows-based server administration, and system hardware.

27 comments
touch0ph
touch0ph

I certainly see some value in offering this service as a maximization of resources. What I'd really like to see is an actual business break down, done independently of course, of say an email system or an application server. What about the added costs of introducing a second ISP as a backup? Or do you just adsorb the risk of Internet access going down? This sounds suspiciously like the SaaS model.

b4real
b4real

You got it, it is the new SaaS - but the meat and bones of what is the offering is hard to get our minds around.

pradhan.ashish
pradhan.ashish

I dont see production-system applications being moved to CLOUD in 2009, however a lot of inhouse investments Eg. IT Test-environments, infrastructure to support Proof-of-concept solutions/Client DEMO,etc maybe the first to go. As the above small-steps bring value and cost-savings to the enterprise and the concept matures.....More will follow. -Ashish

mike_patburgess
mike_patburgess

As one can see from various posts, that there is a lot of confusion stemming from definitions of virtualization, clouds, and so forth. This will not be as simple as defining interfaces like SCSI, Ethernet etc.. From my perspective, you can see no less than a dozen methodologies emerging from this confusion. As a customer, what do you do? Embrace one, spend a lot of money and hope that the one archictecture that you have chosen has the staying power or do you wait until all of the "FOG Coumputing" clears. Personally, this too will come to pass replaced by, the Virtual, Enhanced, Consolidated, Compute, network, storage Engine (VECCNSE).. Catchy, isn't it.

SgtPappy
SgtPappy

I heard of this new great technology coming down the internet pipe. I read that you don't even need a monitor, keyboard and mouse. You just close your eyes and the screen is displayed on the back of your eyelids in high def, no less. You manipulate what you see on "your screen" using thoughts. Talk about speed. You can open any application with literally the blink of an eye. The only draw back would be the occaisional fuzzy display after having lunch with Bud or MaryJane. VECCNSE the next best greatest thing you gotta have before your neighbors.

mike_patburgess
mike_patburgess

Ah you know of the new improved technology. Now that's real virtualization. Virtual computing, open to everyone, free spirits (no not Bud) and of course any worms, trojans, virus' and hackers. As I always say, there is a reason for a business to be in that business which is the profit side of there business and the other is office automation which we could really care less about. The defining line between the two has blurred. The office automation based on a flawed operating system brings with it all of the baggage which we have to protect our business against. Taking this one step further, the business is now running on flawed operating systems instead of a very stable, proprietary OS which for the most part is impervious to intrusion. Wake up people. Proprietary operating systems are not bad. They carry significant benefits. Separate the two business units and protect your investments.

darpoke
darpoke

I'm not sure I see the need to stray from stable OSes. Why would the cloud servers need to run anything else than what a company would use to host those services internally? Our company mailserver, a virtual machine running FreeBSD, is hosted by a third party company - technically a cloud service. We've had it for six or seven years and the only time it's been anywhere near 'unstable' has been one or two isolated incidents. These are cases where one of our staff has tried to send an unfeasibly large file via email, it's been (quite rightly) bounced by the recipient and landed back on our server. Every attempt at mail retrieval after this spawns a new instance of the handler and due to the mailbox being 40 times larger than is acceptable, each of these instances crawls slower than the last until the problem mailbox is located and purged. Naturally, this is a case of PEBCAK and not a fault with the FreeBSD OS or indeed the cloud service we receive.

support
support

Rick: An exact quote from your article says it all for me: "...the fact that we are for the most part working with hype only." I think the cloud is a concept without a clear problem to solve. Once the problem is clearly defined, the solutions can be defined. Until then, it's a lot of marketing hype and your statement quoted above is, I think, accurate. What we need is a unified set of protocols/standards and drop the marketing department's hype at the station while this train gets underway! Thanks!

SgtPappy
SgtPappy

Cloud = Rain No Cloud = Sunshine

darpoke
darpoke

I always thought of Windows as something you needed if you had to work in a box :-) Plus, an Apple a day keeps the (virus scan) doctor away!

markh1289
markh1289

This sounds like the introductory paragraphs to an article ... where is the article with any information in it?

gordon
gordon

In the "Big Picture" perspective Cloud Solutions are analogous to the "ultimate Mainframe". Given the security issues; the possession vacuum; and last but not least the diminished control, the Cloud will succeed because, like the Mainframe, it simplifies the Enterprise. This trade-off in favor of the latter is confirmed by the growth of thin clients signaling the return to Mainframe Architecture. Cloud Solutions are simply the next natural extension of this full circle odyssey. Respectfully, Gordon Haas gordon@ideprize.com

kmdennis
kmdennis

Rick, It seems like this article is dealing with your question here? http://blogs.techrepublic.com.com/networking/?p=1355 What is it you are talking about in this article that is not being covered in your previous article? Or did you forget?

b4real
b4real

I'm trying to get user's take on the cloud. This is more of a forum for everyone to speak up rather than me this time. I'm going to feature the Amazon EC2 cloud next week.

b4real
b4real

We don't have much to go with now. Other than Google Apps. (I'm going to talk about the Amazon EC2 cloud soon). I want to gauge where TechRepublic members are on the cloud in general.

mailboweb
mailboweb

...the advantage here. Even with the servers and the software on the company's side. Will there not be additional cost for using the code of these application. Will Google and or others give the application for free? Yes, its free now for most users, but will that not change when the applications are completed and available as standalone package. If this happens. I don't think company's will be happy when Google database engines can process there data and look for pasterns or other usefulness an probably profitable information. Not to be paranoid, but this looks like an probable effect. They are already doing this. Are they not. The software may be cheaper, but at what expense on the long run.

Deadly Ernest
Deadly Ernest

set up for multi user access, so they need only have a new interface created to allow access over the LAN via a browser. Some of the older mainframe software did word processing and spreadsheets for multiple users over the LAN, they could easily be amended to do the job too.

tweakerxp
tweakerxp

I was thinking about this, what will happen if your company gets waist deep in the "cloud" then the provider decieds that they want to start charging a premium fee for their service? What if all "cloud" providers start charging? Free things usually don't last too long. Just my 2 cents...

darpoke
darpoke

as far as I know, they're one of the only companies offering free cloud services. The concept as it applies to business is a paid-for service, which is exactly how QoS will be enforced. Yes, if it's not stable and reliable enough people will not stick with it. Yes, there is a risk of suppliers arbitrarily raising costs. The same is true of, well, email, telephone services, couriers... pretty much every business-to-business service out there, and to a lesser extent every consumer service too. The need to retain a client base is what keeps companies operating to acceptable levels and deters them from gouging their customers. Contractual obligation plays a large part too. There's no need to rely upon a sense of goodwill.

dshorn
dshorn

The success or failure of this concept will hang on the level of service provided. If something doesn't work, it better be fixed quickly. Whoever is managing the cloud needs to give top flight customer service to the end user. Without that, it will fail.

darpoke
darpoke

I just wanted to add, that of course for most if not all companies this is undesirable, but this business practise is as far as I know largely limited to Google and their particular business model.

darpoke
darpoke

The benefit of virtualisation as a computing paradigm is maximisation of efficiency in utilising resources. When a server is running a simple or sporadic process, the capability of the CPU(s) and other hardware attributes go largely to waste. One solution is to group services onto smaller numbers of servers. There may be security issues making it unwise to cluster specific resources together on the same 'machine'; virtualisation negates the need to do this*. Timesharing is another benefit of virtualisation. Services that require reciprocal quantities of resources at different times of the day are able to run as virtual machines on the same server. This can be achieved without virtualisation of course but the changeover is far more seamless and swift when using virtual instances instead. In short, to cluster the various services needed/provided by your company as multiple virtual instances residing on a single machine of sufficient capability allows a department or company to shave the excess computing potential from all hardware at once, while aggregating them together may enable tessellation that permits even further savings. The impact this has on staffing needs can mean fewer trained personnel are required to maintain the same number of processes; since some of all maintenance consists of hardware service and troubleshooting, reduction in the quantity of hardware simplifies the matter. In some cases the geographic accumulation of devices in one area can also aid maintenance. To observe the benefits of cloud computing, consider the above paradigm, as applied instead to multiple entire IT departments. If a third party company (whose only business is running virtual services for other companies) has multiple clients, then they can aggregate their virtual services and benefit from them to a similar degree that a single company adopting virtualisation can. A 'cloud computing' company operating a large server farm can benefit from centralised hardware, the same impact on staffing as detailed above, and the potential for saving through bulk purchasing or licensing thanks to economies of scale. This can lead to low enough running costs to permit provision of the service that still beats the cost to the company of running such services themselves, if one includes the cost of sufficiently trained staff, a permanent redundant backup system, uninterruptible power sources and so on. *While virtual servers naturally share the same hardware, the potential security risks of having critical resources on the same server are substantially higher than those of having such resources on two virtual machines co-located on the same physical device.

jeff_oconnell
jeff_oconnell

I'm in total agreement with darpoke, but I think what he's saying needs a little expanding on.? Cloud computing? means different things to different people, and because of this the "industry" needs to do a better job in explaining just exactly what they are talking about. Using today?s understanding of "cloud computing" although, it's accurate to say that Gmail is to date the best example of a successful cloud computing concept it's also inaccurate. The reason being Gmail is a hosted email provider, and hosted email providers have been around forever so why is Gmail the best example. The reason is simple. What Gmail is really trying to do is push forward is not a change in technology, but a change in concept. They are shifting the focus away from "software packages" to "service providers". Forget about MS Office ?the application?, think about Google Docs "the service". This is what I believe darpoke is getting at, but it?s being lost in translation. "Cloud computing" is really about divorcing IT from back-end infrastructure and refocusing on delivering IT services to the end-user cheaply, and efficiently. By that I mean providing the end-customer the ability to take their "server" and port it from one company to the next at a moment?s notice. The same way companies change hosted web-server providers. Giving the customer the ability to add and subtract resources as needed and not through bulk-buy, because he/she wants the cheaper rate, and hope it's utilize it down the road. And to give the Infrastructure companies the ability to better maximize their investment. Let?s not forget ISP?s make very little to no money at all. That is why contention ratios are so friggin high and SLA?s response times horrific at best. It?s not about operating systems, Microsoft versus Linux, MS Office versus Open Office, or IT security, etc. Truth is virtualization is most likely going to bring about OS agnosticism. Cloud computing is here today. What's stopping it is mostly the SMB. SMB's aren't going to buy into something they don't understand. Especially, when they have no real justification for the expense of conversion when what they have is perfectly adequate for their needs, even though adequate may not be very good. The industry needs to do a better in providing a simplistic explanation that everybody can understand.

darpoke
darpoke

It's great to have someone come at it from the other side. You covered loads of aspects I managed to miss entirely. Of course the technology behind cloud computing is nothing new. Our company has had its incoming mail hosted by a virtual server running an install of FreeBSD for six or seven years now. We've not experienced a single problem that wasn't caused by user error (I made the mistake of teaching my co-workers what PEBCAK means, but at least now we can all laugh about it good-naturedly :-) ). I think the benefits to be had are unified service across multiple sites for companies where this is an issue; cost benefits to be had from a company that can buy servers far more cheaply than you can, while retaining staff you don't have to pay full-time wages for when maintenance needs are often sporadic at best; and potential off-site redundancy of information backup. In re-reading Jason's article, I realised that while he discussed why current barriers to adopting 'cloud services' may well disappear or at least fade into the background in the near future, he didn't present any of the arguments for adopting it in the first place. I suppose this may have been the central point of other articles on the subject, and the barriers were the focus of this particular piece. The truth is, for a company with an established IT contingent, the benefits are few or none. The move to cloud services for such a company would involve liquidation of equipment and dismissal of staff. The arguments for cloud services in this scenario are simply arguments for virtualisation if it is not already in practice. It becomes more attractive to companies looking to start up, or expand, and would need considerable initial investment in staff and equipment. The granularity offered by cloud services means companies are free to utilise only what they need, avoiding paying for excess computing or staff whose roles will taper off or subside once the service has been established and stabilised. The ability to dial the level of service up and down as necessary is also one that internal IT paradigms often lack. There's little to be saved once servers have been invested in, aside from switching off the superfluous machines and saving on the running costs.

tom.marsh
tom.marsh

Virtualizing servers to consolidate hardware is a brilliant strategy. In some cases, it is also a brilliant strategy to outsource IT-service provision to somebody else outside the company. These things make sense. These steps simplify management and reduce costs respectively. For sake of argument, take your e-mail and calendaring and switch to Google Apps. Its pretty sweet. Now imagine the neighbor has a crew digging on his property and breaks your fiber. Snap. No more Internet. ...and it could be down for DAYS depending on SLAs, the relationship between the entities that manage the fiber and the entities that manage the IP network flowing over it. "Cloud" computing leans heavily on this vulnerable stop in the path network traffic takes from your site to the remote data-center. A second circuit might help, but unless you have it writing that the second circuit takes a different path than the first one (doesn't ride the same fiber, preferably takes a different physical route too in a different conduit) you haven't really increased your redundancy, just the speed at which your employees access Facebook. ...and this is a pretty big "gotcha." In the past, an "internet" outage meant you could email people at your company but not outsiders. Work could proceed, albeit with outside contact being over the phone. Now imagine what happens if you've moved your productivity apps off-site. A two-hour "oopsy" with your Internet service kills 25% of that day's productivity with email and calendaring.

blarman
blarman

The biggest issue for many businesses will be to see if cloud-based versions of applications can successfully mimic individualized services to businesses. Every business has their way of doing things, and the software they use reflects the way they do business. The biggest obstacle to the cloud is going to be the customization of their products to meet the needs of the individual customer. The cloud is a nice idea for those with big fat Internet connections, but the even larger concern will be the product itself and how it matches the customer's business practices.

b4real
b4real

Because virtualization among the more 'textbook' on-premise cloud solutions that we have to date.