VMware Releases Open Source Cloud Foundry

Julie188 writes "VMware shook the cloud world with an announcement that it was releasing an open source platform-as-a-service called Cloud Foundry. Not surprisingly, the new cloud platform takes direct aim at Microsoft's Azure and Google's Google Apps platforms. Cloud Foundry is made up of several technologies and products that VMware has acquired over the recent past and is released under an Apache 2 license. While VMware isn't the first-and-only player to launch an open source cloud initiative (Red Hat has DeltaCloud, Rackspace and Dell have OpenStack), some believe that with VMware now in the open source cloud business, pressure could be mounting for Microsoft and Google to release versions of their cloud that could be hosted somewhere other than their own data centers."

Comments

Public/Private/Hybrid

[Jeremiah Cornelius]

Run your whole public cloud infrastructure and application fabric on the same technology platform as you use to manage your internal data centre.

This is a better by far option than Microsoft - who's idea is to land an Azure container at your doorstep. And it scales from the tiny to the gigantic.

The heart of this stack seems to be gold old Tomcat. The path to an application layer that is aware of on-demand elasticity seems very good.

Press release

[Nimey]

The summary reads like a press release. "Shook the cloud world", indeed.

Re:I must be old but...

[Anonymous Coward]

there are benefits to not owning your own servers. Cost being the biggest one.

Less labor costs, health care benefits, and retirement packages up front because someone else hires the monkeys to maintain the hardware. If you are a corporation, you have to pay taxes on all assets owned by the company. Not owning the hardware means less taxes to pay on assets. Less hardware also means less real estate you have to own to house that hardware, which in turn means less real estate taxes.

the savings goes on and on...electricity, water, HVAC equipment, etc... If the cost benefit is big enough, risk of handing over the keys becomes less of an issue.

So let me get this straight...

[GuruBuckaroo]

We move all our department stuff into the Data Center. Consolidate all of our equipment into one spot, less field work, fewer techs.

We move our Data Center into the Cloud. Less equipment in our Data Center, even fewer techs and admins, reclaim power cooling and space.

Now we move the Cloud back into our Data Center? What's next, distribute our data center into the branches so it's disaster-tolerant?

... wait...

Re:So let me get this straight...

[rosciol]

Not quite. This move is all about options and preventing the dreaded 'lockin'.

Developers ask: If I bet my development on a single PaaS provider, won't I be tied to them indefinitely?
VMware says: If you use our open source stack that can be hosted by us, or by you, you won't be tied to a particular framework or hosting provider. We'll happily host you if our service fits your needs but, if your needs outgrow us or we fail to meet your quality expectations, you can always run the exact same stack out of your own datacenter or someone else's.

What remains to be seen is how good the performance is and how easy it is to use the platform.

Re:So let me get this straight...

[Rich0]

Well, a big thing in cloud providers is that there is no red tape in provisioning. I go to EC2, fill out a one-time form to set up an account (about as complicated as buying a pair of shoes on the site), and I just pick an image and tell it how many servers of what size to fire up, and in a few minutes I have a list of DNS addresses to ssh into.

If I could have that at work it would be wonderful. Right now getting even a new virtual server seems to take forever. I'm all for resource accountability, but there is no reason provisioning should be so manual.

I'd love to see something like this at work. And, of course it gives you the ability to send peak demand offsite easily.

Yawn

[ugen]

When is this "cloud everything" fad going to be over? It's a data center that someone else runs for you. Big deal. (Sure, when you put it that way - it does not sound nearly as cool and does not sell so well, does it)

Re:Yawn

[subreality]

It's a data center that someone else runs for you. Big deal.

Said like someone who's never had to deal with a data center before. In the small to mid-size business range, running a data center seriously sucks. It involves at least one of: talent and capital that small businesses just don't have; renting a cage from someone who's charging an arm and a leg for mediocre service; or simply building a poor one in what used to be a conference room and dealing with crap power, cooling, cabling, etc, because those hassles are still cheaper than either of the two prior options.

Hassle-free sign up for what you need as you need it with no lock-in is a SMB's dream come true. I don't think it's going away.

Re:Yawn

[seifried]

For now, centralizing computing resources in "the cloud" doesn't have the same obvious and conclusive upper hand in the vast majority of common uses that centralized electricity generation does.

It's rapidly happening. You want to do some computing, you can either build a compute farm/etc. (and a server room/building, cooling, electrical, etc, etc.) or just run it on EC2 (which means setting up some non trivial software to handle firing up all the images, monitoring them and their results, setting up storage, etc.), or better yet outsource the whole mess to Cycle Computing (which in turn uses EC2, but handles all the setup/etc. which is non trivial). Compare Cycle Computing at $1,060 an hour (for 10,000 compute cores) and they handle EVERYTHING vs. doing it yourself. http://www.theregister.co.uk/2011/04/06/cycle_computing_hpc_cloud/. Plus you need to remember tax wise operational expenditures are often preferred over capital expenditures, and if your compute farm isn't computing it's a waste of hardware, whereas on EC2 you can provision a ton of systems for the hours or days you need them and then turn them off and stop paying.

Re:uhmm

[BitZtream]

why was Amazon not named in the summary?

Cause its more of a place to get access to extra CPU power cheaply for a period of time rather than a place to host services as their primary location since it is ridiculously overpriced compared to ... well everyone else? AWS may get used to host some services in their entirety for some companies other than Amazon, but not any smart ones. AWS is where you go when you want to run a massive set of number crunching or processing, or do a big release of something and utilize there content distribution network for a couple weeks until things get back to a level at which your standard hosting environment or datacenter can handle.

Who cares about Microsoft

Anyone more concerned about what using the right tools for the job at hand rather than being an ignorant flaming fanboy?

and who knew Google had a cloud offering.

Anyone who has been paying attention 'cloud computing' anytime within the last couple of years? And a ton of other developers who happen to use those cloud services for their own projects (rietveld, not google code), and all the companies who sell products running on Google's cloud (of which there are many) and all the companies which utilize products sold by companies using Google App Engine (of which there are more that probably anyone else other than AWS).

Perhaps you should get a Cloud Clue before talking about it?

Re:I must be old but...

[Anonymous Coward]

I'm not sure if someone mentions it later, but the cost structure is an even bigger driver than most people realize. If you buy the servers and the software for them, that's classified as a capitalization expenditure, which means that only the depreciation of the asset can be deducted over time (most IT resources is 4 years). However, if you get the same thing as a service, it's fully deductible every year. Which means that cost-wise for the equipment, instead of a large upfront cost with many years to recoup the difference, you're getting it up front. For cash-strapped startups or even small and medium businesses, that's a huge deal.

Another benefit for publicly-traded companies in the US is SOX-compliance. Move the data/services to the cloud and the data-retention issue is now the provider's problem, not the company's.

Re:I must be old but...

[cduffy]

You call them up they spawn a VM and send it through the billing app once a month.

Huh? Phone a human to spawn new machines? Monthly billing? That's missing the point.

Billing resolution down to the minute, on the other hand, with fully automated management is actually taking good advantage of what's available -- if your load is spiky, maybe it's worth the markup to launch new machines based on current load (or pull underutilized systems out of the pool and shut them down) and pay for only what you use. ...and by the way, if you think you can hire a full-time UNIX admin and buy and host a rack full of 2Us for less than you'd pay a competitive cloud vendor (note, I said "competitive", not Amazon)... well, your idea of the numbers is waaaay off. I've played it both ways, and next time I'm doing a ground-floor startup, the last place I'll want to be is managing physical infrastructure.

Re:Op Ex. versus Cap Ex

[TheRaven64]

There's also the advantage of specialisation. Hosting 1 server is expensive. Hosting 10 servers costs about the same amount. Hosting 1,000 servers costs a lot less per server than hosting 10. There's no way that I could even get an Internet connection with the kind of throughput that a typical colo company offers for even the price of hosting, and that's ignoring the cost of power, the cost of the space, and so on. With a VPS it's even bigger. A lot of small companies only need 10% of a machine, but hosting that themselves would cost much more than renting it.

At some point, it becomes cheaper to move it in house. If you're a big company, then it definitely makes a lot of sense to host your own infrastructure.