ServersThe Problem with Private Clouds

The Problem with Private Clouds

ServerWatch content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.




Private clouds are great for companies that can’t (or won’t) use a public cloud. There’s just one problem.

The problem is that simple-to-implement private cloud solutions don’t come cheap (think VMware), and cheaper solutions aren’t simple to implement (think OpenStack.) Virtually Speaking

That’s just the way cloud computing works, and it’s a bit of a problem for a large number of sub-enterprise organizations that lack both financial resources and OpenStack skills.

It’s this type of organization that’s the target market of ZeroStack, a California-based company that has just launched its ZeroStack Cloud Platform, a private cloud-in-a-box solution the company touts as “zero touch.”

Actually “cloud in a box” is not quite accurate. The platform is supplied as one or more appliances that integrate compute, storage and networking, and which are preloaded with the OpenStack-based private cloud software. It’s on this hardware that the organization’s apps, data and ZeroStack management software reside.

But the platform also has a cloud component too, which administrators log in to in order to carry out activities such as spinning up VM images, allocating usage quotas, carrying out chargebacks and performing analytics to provide “insights and optimizations” for application deployment on the ZeroStack appliances.

The key selling point of the platform is that you can plug the boxes in, switch them on, and your private cloud can be up and running in less than thirty minutes, the company claims. This contrasts pretty favorably with typical OpenStack implementations that can take weeks or even months to get off the ground.

Nick Suh, ZeroStack’s head of marketing, offers the example of Florida Atlantic University (FAU), which has already installed a ZeroStack private cloud. “We shipped them an appliance, and they connected it and gave it an IP address,” Suh says. “After that we could discover it in our cloud resource and help them configure it. They were up and running in minutes, and we have never been to their [physical] site.”

He adds that after a system is up and running, administrators need to configure network policies and security settings. “A basic setup takes about half an hour to an hour,” Suh says.

The ZeroStack system comes with predefined templates for popular open source applications such as Hadoop, Jenkins and various OS images that are available from a self-service catalog. Business units and users can then help themselves to the templates with a single click – once they have been granted permission to do so by an administrator.

In terms of cost, Suh says ZeroStack is much cheaper than a VMWare private cloud. “FAU looked at VMware and its vCloud Suite, but it couldn’t make the economics work,” Suh says. “Relative to VMware we are one-third of the cost looking at three-year TCO.”

(While that certainly sounds impressive, it is of course always wise to treat vendors’ TCO comparisons with caution and do your own math to ensure that the time periods and assumptions made by the vendor match your own.)

A Variety of Target Markets and Use Cases for ZeroStack

We touched on target markets for ZeroStack earlier, but Suh suggests some more specific examples. “Companies that want a private cloud but don’t want to employ more than five OpenStack engineers. Test and development would be another use case. Startups, midsize companies, LOB of enterprises, any company doing software development or lots of number-crunching and analytics” would be other ideal scenarios for ZeroStack, according to Suh.

He also suggests companies where “shadow IT” has been using resources in a public cloud and where there’s a need to run those resources privately for compliance reasons, as well as companies that do want to use the public cloud eventually, but wish to start in-house.

These last two use cases are made simpler because ZeroStack supports Amazon Machine Images (AMIs). That means they can easily be brought from AWS to a data center running ZeroStack.

Likewise, it’s possible to start on ZeroStack and then migrate to AWS (or Azure or Google’s cloud for that matter). It’s also possible to operate a hybrid cloud using ZeroStack and any of these public clouds.

In technical terms, ZeroStack’s appliances are delivered as a 2U chassis with four sever nodes, with 64-128 CPU cores and 512GB – 1TB RAM. Storage is proved by 6.4 – 19.2 TB SSD storage and 16 – 92TB hard disk storage. (Different versions are available offering more compute power or memory and so on to suit different sets of applications.)

Additional appliances can be added to scale the system to hundreds of nodes, and a software-managed distributed, self-healing architecture eliminates the need for IT administrator involvement to maintain a resilient environment, according to the company.

In terms of cost, a bundle including an appliance, software, access to the cloud portal and support costs up to $100,000 per year, Suh says.

The ZeroStack platform is generally available as of today, March 8th, 2016.

Paul Rubens is a technology journalist and contributor to ServerWatch, EnterpriseNetworkingPlanet and EnterpriseMobileToday. He has also covered technology for international newspapers and magazines including The Economist and The Financial Times since 1991.

Follow ServerWatch on Twitter and on Facebook

Get the Free Newsletter!

Subscribe to Daily Tech Insider for top news, trends & analysis

Latest Posts

Related Stories