The Wisdom of Clouds - Latest Comments

Re: The enterprise "barrier-to-exit" to cloud computing

Maisa — Tue, 02 Dec 2008 12:07:11 -0000

Interesting. Definitely the ones after better resources for their move into the cloud will be more prepared and assisted, with less possibilities of dealing with future security flaws or even, as you well put, high costs to do it. And even this article being extremely centered on companies, it's also important to notice the final users move into the cloud: will eventually Web OS serve the final user just as in good terms as companies services? will final users have their space and security for their own data?

Re: The enterprise "barrier-to-exit" to cloud computing

TomBasham — Tue, 02 Dec 2008 08:54:45 -0000

Interesting post, I think you're right, existing investments are definitely a barrier to wholesale migrations to a cloud model. This does however provide a migration path for those organisations that are investigating cloud options for the provision of services.

At the moment there aren't too many services available from the cloud in a cost effective way - at least from the big players that large enterprises will look to deal with. At the moment we're finding that our desire for cloud type solutions far outweighs the markets ability to deliver them. As existing on-premise services start to age I think we'll see a service-by-service migration to those cloud services that are mature and have a cost effective billing model. For some services this will work out well, but for others the market won't be ready and the services will enter another life/depreciation cycle as on-premise solutions. It's great to see VMWare et al creating solutions that will help enterprises build their solutions in a way that will support and enable cloud computing in future.

Tom (twitter.com/tombasham)

Re: What is the value of IT convenience?

jamesurquhart — Tue, 02 Dec 2008 06:30:24 -0000

Well said, Steve. And I completely agree that IT is now under pressure to compete, not just serve. Let's see how the industry reacts, shall we? :-)

Your company/product naming comment is noted, and I made the correction.

Re: What is the value of IT convenience?

Steve Caughey — Mon, 01 Dec 2008 13:20:16 -0000

James,

Billy Marshall’s comment that IT departments are being sidestepped is something I’ve heard on many occasions now. IT departments need to wake up to the fact that they’re competing, directly and increasingly, with public clouds. They need to get their costs down and agility up (and I suspect that they’ll only be able to achieve that by themselves utilising public clouds as part of their back-end infrastructure). They can then focus on delivering value-add through knowledge of the particular concerns of their internal customers. IT departments need to turn themselves into the cloud of choice for those customers.

I’d also agree with his comment on how cloud infrastructures can increase demand. I think the equation is even simpler than the one he uses. If we decrease costs (through increased sharing and economies of scale) demand will increase. Period. This extra demand will not only come from new applications which become viable with the use of on-demand cloud resources, but also from applications which are capable of delivering more value but which are currently constrained by the cost of utilising extra compute resource. These include complex simulations (financial, engineering and scientific) and data mining (geophysical, petrochemical, pharma etc). If the costs of running those applications decrease then, in many cases, we’ll just run them more.

Incidentally, thanks for the mention, though it’s worth pointing out that our product name is ‘Agility’ but company name is ‘Arjuna’.

Re: What is the value of IT convenience?

BenjaminTseng — Sun, 30 Nov 2008 02:11:10 -0000

Is it a bad or even unnatural that we're seeing this explosion? Given the relative novelty of these "cloud" techniques, one would imagine that IT is in the midst of a transition, with inexperienced IT departments inefficiently allocating resources today as they are unfamiliar with the new offerings, and nor are the service providers particularly experienced at pricing or setting their offerings. This short run "chaos" will probably eventually settle down when IT managers re-rationalize their IT infrastructures and cloud service providers rationalize their offerings.

Re: Do Your Cloud Applications Need To Be Elastic?

jmason — Wed, 26 Nov 2008 10:18:58 -0000

Hey James -- sorry for the delay, disqus didn't let me know there was a reply ;)

We're an early-stage startup, so no, we had no existing infrastructure bar a few colo servers which could be entirely replaced by EC2 instances.

Being able to use small numbers of servers at a time and be billed for just the CPU time used is great for us. Of course, we also have inelastic parts of the infrastructure that could be hosted elsewhere at a colo for less cost, and personally, I would probably have done this given the choice; but mgmt were happier just to use EC2 as widely as possible, despite the additional costs, since it keeps things simpler.

For example, we also use S3 heavily, and EC2-to-S3 traffic is extremely fast and cheap compared to external-to-S3; so that's proving to be an effective point of lock-in. (We don't mind.)

Regarding locally-hosted servers on our own infrastructure -- that was simply out. There's a massive amount of investment that would be required to get a sufficiently-reliable infrastructure set up, and a good deal of coding and config to get an EC2ish elastic server provisioning system going on that. Further down the line, that might be appropriate, but not yet... it's a lot easier to let Amazon do it for us ;)

There's definitely a point where we *could* migrate to our own infrastructure in the future, and open source apps like Eucalyptus make that more viable -- but in my opinion it's very far off indeed...

Re: Do Your Cloud Applications Need To Be Elastic?

jamesurquhart — Tue, 25 Nov 2008 15:05:04 -0000

See, now that's an interesting use case.

Did you do a cost comparison compared to running virtual machines on your own infrastructure? Did you have your own infrastructure to start with?

Also, have you ever done a spreadsheet projection about your costs if you become wildly successful? Is there a tipping point where owning your own stuff starts to look more attractive? That's what Animoto was predicting a few months ago.

Thanks for the comment. I hadn't thought of this angle at all.

James

Re: Do Your Cloud Applications Need To Be Elastic?

Justin Mason — Tue, 25 Nov 2008 11:08:57 -0000

There's another reason to use an EC2-like cloud without horizontally scaling a single application; if you want to deploy another copy of the app, either from a different version-control branch (dev vs staging vs production deployments), or to run separate apps with customizations for different customers. These aren't scaling an existing app up, they're creating new copies of the app, and EC2 works nicely to do this -- I'm using it for that purpose right now.

(Also, w.r.t. figures: most of our EC2 nodes are the 10c/hour variety; the classic "many cheap servers vs. fewer big servers" scaling model.)

Re: Is IBM the utlimate authority about cloud computing?

Bud Gibson — Tue, 25 Nov 2008 02:41:11 -0000

I have to confess you are infinitely more on top of this topic than me, but my question is this: Will anyone actually look to IBM for advice on their cloud computing infrastructure? Well, maybe some enterprises. But, is that the market players like Amazon are trying to pursue? Everything I see on Amazon tells me web startup or short burst compute project.

That's the actual market that seems to me to exist today for cloud computing.

I don't think anyone has a real grip on how to bridge between traditional enterprise applications and the cloud. Oracle's Amazon Machine Instance is a case in point. I can't buy an hour by hour license for that. I've got to pony up for the full thing, sort of defeats the purpose.

Re: Is IBM the utlimate authority about cloud computing?

EthanBauley — Tue, 25 Nov 2008 02:07:28 -0000

Have you seen the IBM virtualization conversation?

www.VirtualizationConversation.com

Re: Do Your Cloud Applications Need To Be Elastic?

jamesurquhart — Sun, 23 Nov 2008 16:03:28 -0000

No, but then the Amazon figures don't reflect administrative OPEX, security/patching OPEX, regulatory compliance OPEX, backup system OPEX (or possibly even CAPEX), et. al. Don't be fooled, Amazon just gives you some servers; you are still responsible for provisioning, monitoring, patching, securiting, assuring compliance, providing a backup plan should Amazon go down, etc., etc., etc. Developers think there is nothing to do once the server is provisioned. System administrators responsible for availability and compliance know much more is at stake.

Re: Do Your Cloud Applications Need To Be Elastic?

Roland Dobbins — Sun, 23 Nov 2008 14:59:55 -0000

From whence are these figures derived? Do the on-premise figures reflect administrative OPEX, security/patching OPEX, regulatory compiance OPEX, backup system CAPEX/OPEX, et. al.?

Re: Do Your Cloud Applications Need To Be Elastic?

jamesurquhart — Sun, 23 Nov 2008 08:37:03 -0000

Again, if it doesn't need to be running all of the time (and you can remember to shut the dang thing off), it may be a candidate for EC2/S3. However, just note that there are 24X7 costs for every EC2 instance, whether it is on or not, in the form of S3/EBS.

I'm just warning people to do the math before committing to the cloud.

Re: Do Your Cloud Applications Need To Be Elastic?

james — Sun, 23 Nov 2008 05:41:05 -0000

But how do you spin up a program for 12 hours a day on EC2 without having a presence already in the cloud that you would have to pay for?

And these number don't include maintenance, bandwidth, and S3. Keeping a data center up is not a part time job.

Re: Do Your Cloud Applications Need To Be Elastic?

Tom Gleeson — Sat, 22 Nov 2008 13:56:37 -0000

Even in the case of accounting systems, CRM etc., many business work less than 12 hours per day, 5 days a week, if the instances are powered up only for those hours of use(as many businesses currently do with their in-office servers and PCs), then the cloud can also be cost-effective for these types of applications.

Tom

Re: Reuven Cohen Invents The "Unsession"

Reuven Cohen — Fri, 21 Nov 2008 18:08:15 -0000

I aim to please. I actually did do a full hour long power point slide for the session. But after seeing the other vendor centric pitches early that day, I figured it was time to do something unconventional. Thanks for the kind words.

Re: Why the Choice of Cloud Computing Type May Depend On Who's Buying

Alain_Yap — Wed, 19 Nov 2008 09:03:45 -0000

Just wanted to share James that we get inquiries into tweaking the Platform as well for something that we don't provide support yet in addition to integration concerns from existing companies. I take it that PaaS (like what Geva mentioned) will go on the direction of being multi-flavored while incorporating customization up to a point.

Regarding IAAS, I guess it will be more of a decision that will be more affected by financial considerations more than any other factor. It may just be me but I see IaaS players creeping features into their products that somehow mirror something like PaaS in the long run. [Gray]

Best.

Re: "Follow the law" computing

Sjs information — Wed, 19 Nov 2008 07:18:41 -0000

As far as i think...the isue you have included here is very real & true.It was quite informational. I personally liked the Sharding/Replication part , the most.But i think that laws are made to be followed & so should they be.

Re: "Follow the Law" Meme Hits the Big Time

Sjs information — Wed, 19 Nov 2008 07:13:13 -0000

I too agree that Nick's Big switch is an eye opening post.Thanks a ton for specifying it here.I really liked reading it.

Re: Why the Choice of Cloud Computing Type May Depend On Who's Buying

jamesurquhart — Mon, 17 Nov 2008 06:44:00 -0000

I think both Geva and Stuart are correct, but the differences come out when you dig into near term versus long term. In the 3-5 year horizon, there is no doubt that IT organizations will use the models they know and love when leveraging the cloud, in all its various forms. So, IaaS is attractive for those wanting server capacity for software stacks they own and operate. Specific types of developers developing specific applications will be drawn to PaaS, however, as it eliminated gaining a skill they'd rather not waste time on.

However, my question stands for the 5-10 year time frame. If the ultimate goal from the businesses standpoint is specific functionality at specific service levels (including security, etc.), then why "build you own" if someone offers a perfectly good framework/platform on which to build it.

Now, folks like ELASTRA, Cassatt, 3TERA and others have an opportunity to drive upwards from infrastructure automation into "platforms", but I predict they will have to narrow down their target market to do so. Alternatively, they can remain general, but will become "plumbing" and relatively commoditized if they do.

I just have trouble seeing a CIO agree that hiring an entire data center staff beats just hiring a developer to write functionality on an infrastructure that someone else worries about--with the caveat that ALL of the CIOs biggest concerns would have to be met before PaaS would be selected. No PaaS platform is there today, but I'm pretty sure that if the money is there, there will be several that target specific business uses in the mid to long term.

Re: Why the Choice of Cloud Computing Type May Depend On Who's Buying

Stuart Charlton — Mon, 17 Nov 2008 03:07:53 -0000

I agree with the categories, but I look at the results differently.

Think of how either model enables an ecosystem that parallels the rest of the IT industry (both individuals and/or companies). The Fabric approach seems to be an "all your base are belong to us" strategy, placing the developer at the centre of the supply-side of the IT industry ecosystem, whereas the Architect, SysAdmin, DBA, etc. are "taken care of". The demand-side targets only the silo'd application consumer (in the case of SalesForce's AppExchange), but doesn't really target the IT manager that's worried about application integration, data quality, security, etc.

The Instance approach is much more inline with today's industry ecosystem -- it primarily leads to a more efficient (lower lead time) supplier & parts network. There still are independent software vendors. And architects. And sysadmins. And IT Managers that prefer to "trust, but verify" their SLAs. In this world, systems are more "visible", and have much better lead times than the traditional data centre.

In-between these two categories, I might add, are the Cloud "Glue Players" like 3Tera, RightScale, Elastra, etc. that are trying to raise the bar on Infrastructure, adding in configuration management, integration, autonomics, etc.

Which will win? Well, on the public web, developers tend to have a lot of sway, so "Fabric" seems to play well for some crowds, except it represents a severe form of lock-in. In the case where the "locked car trunk" is well upholstered and available in both private & public clouds, such as MIcrosoft's .NET and Azure, I can see success.

But in the enterprise, developers typically have very little sway over data centre investments. They may pick software technology (some of the time), but there are enterprise architects that deal with strategic vendor relations, and there are data centre / operations folks that deal with keeping the lights on, and all three are from very different worlds, with very different assumptions.

Being with Elastra of course, I'm biased, but given my above assumptions, I think both Fabric and Platforms have room to coexist, Fabric slanted towards Silo applications for SMBs and custom-apps for Web shops, with instances (and some kind of "Cloud Glue") being the preferred choice of Enterprises for at least the next 5+ years.

Re: Why the Choice of Cloud Computing Type May Depend On Who's Buying

Geva Perry — Sun, 16 Nov 2008 17:35:59 -0000

James - Excellent analysis. Although you kind of allude to it, I would add that because PaaS tries to hide some of the underlying complexities, it tends to be packaged in a way that suite a very specific application type. For example, Force.com is highly suited for database-centric apps, but is perhaps not ideal from event-driven architectures. Google App Engine seems to have been designed with Web 2.0 style web apps in mind. My personal opinion is that over time we will see more of these PaaS platforms come in different flavors and the need for pure IaaS platforms will be reduced (although never eliminated).

Re: http://blog.jamesurquhart.com/2008/11/in-cloud-computing-good-network-gives.html

Howard Oliver — Fri, 14 Nov 2008 19:24:35 -0000

Hi James: You might want to take a look at EsoteraS3 that has a cloud-based storage solutiion. Their CTO and President Tom Chalker would be an interesting person for you to talk to. He can add to the ideas you developed on storage allocations.

Re: A Quick Guide To The "Big Four" Cloud Offerings

Jim Salem — Fri, 14 Nov 2008 17:12:50 -0000

Nice basic summary. As an architect on Intuit QuickBase and Workplace.intuit.com, I'd suggest we most closely fall into the SF category. We're trying to make it very easy for devs to focus on their apps without the bother of having to maintain an entire VM (ala AWS).
Another summary I like that's focused on just the DB components of these products is here: http://oakleafblog.blogspot.com/2008/04/compari...

Re: http://blog.jamesurquhart.com/2008/11/in-cloud-computing-good-network-gives.html

Greg Ness — Fri, 14 Nov 2008 01:05:04 -0000

Thats one of hte reasons I think that both virtualization and cloud will be big for the network players.

Greg (Archimedius)