Nebula, Cosmos, And Private Cloud's Future

Translated, the Latin word "nebula" means "cloud," appropriate for a company founded by former NASA CTO Chris C. Kemp. Last week I sat down with Kemp, founder and current chief strategy officer of Nebula and one of the inventors of OpenStack.

Now, let me state upfront that I've never been a fan of companies packaging and selling open source projects. So when I first heard about Nebula, I had a presupposed opinion that I would not like it. I was wrong -- Nebula has definitely impressed me. Maybe the gas and dust that make up the actual nebula cloud are having an effect?

Nebula sells hardware and software meant to help companies set up secure, OpenStack-based private clouds, without needing OpenStack experts.

With all the buzz around OpenStack, people sometimes forget how difficult it is to install, maintain, and most importantly secure. Nebula's Cloud Controller hardware appliance runs the company's Cosmos distributed cloud operating system. Cosmos can quickly assemble standard x86 servers from different manufacturers into nodes in a large, distributed private cloud.

Nebula recognizes that one downside of OpenStack is the difficulty in maintaining it, especially from a version and features perspective. So it ships its Nebula One system with a managed service that abstracts the update and upgrade process. With an entry price of less than $100,000, it's enticing for enterprises that want to stand up private clouds quickly but lack OpenStack expertise. I recently spoke with Chris Kemp, founder and current chief strategy officer of Nebula and one of the inventors of OpenStack

Elias Khasner: How do you summarize Nebula?

Chris Kemp: Nebula is a turnkey private cloud appliance. We take OpenStack, the leading open source cloud platform, with significant backing and contribution from anyone who is anyone in the technology world, and package it into a simple plug-and-play deployment that enables organizations to build on-premises private clouds that are very similar to Amazon Web Services.

EK: There's a lot of buzz about OpenStack, but I am not seeing enough enterprise adoption. Would you attribute that to the rapid development cycle?

CK: On the contrary, I think the fast development cycle that the OpenStack project is displaying is a healthy sign. I would even venture to say that there is no such thing as a development cycle [that's] too fast in cloud. We welcome the active development and progress of OpenStack. As fast as OpenStack is going, it's still hard to keep up with public cloud "pure-plays" like Amazon, Azure, and Google.

Enterprises are adopting products powered by OpenStack, like Nebula, today. I don't expect most enterprises to adopt OpenStack directly. People that perceive OpenStack's rapid development as a challenge are those trying to implement it in their products and services. This is the big misconception -- OpenStack should be viewed as a toolkit, not a finished product. OpenStack will never be finished, nor will it ever be optimized for enterprise deployment, integration, or operation. The way enterprises will adopt OpenStack is by turning to companies like Nebula that have built OpenStack powered products that are designed to be deployed into, integrated with, and supported by enterprise IT organizations.

EK: Are you saying that enterprises cannot implement OpenStack?

CK: Not exactly; the project is available for anyone to use. But for organizations to properly deploy it, they will need an army of developers and consultants that will cost them hundreds of thousands of dollars (or more) and will not be completed in any reasonable amount of time. I can say without reservation that most organizations that have tried to harness OpenStack to build an internal private cloud service have arrived at this conclusion. Securing OpenStack alone is a serious undertaking.

EK: How does Nebula address the security concerns in OpenStack?

CK: Nebula has invested heavily in ensuring that our product is secure by design. Securing OpenStack is definitely the hardest and most complicated task of all. This is one of the beauties of adopting a technology like Nebula.

EK: What do you think about the rumored VMware "Project Mystic"? I can't help but see similarities between Nebula and what VMware will announce.

CK: If the rumors are true, then this validates our approach, and if VMware (or Microsoft) can't build a more simplified "private cloud" deployment of their platforms after spending billions of dollars on a proprietary, vertically integrated software stack, then who can?

EK: Is Nebula a replacement for VMware?

CK: Not at all, and that is a wide misconception. VMware's purpose is to virtualize existing workloads that rely on the infrastructure to be highly available and perform[ing]. VMware is optimal for the majority of software that is not infrastructure-aware at all. Public and private cloud are optimal when applications can be made more reliable or perform[ing] [because] they can control the infrastructure directly through APIs. Nebula, Amazon, and IaaS in general assume a service-oriented architecture, where the application is aware of the infrastructure and can self-manage its own availability. [These apps] are the perfect candidate for Nebula and the cloud.

One of the advantages of Nebula is accelerating the pace of private cloud build-out, but more importantly, they force the enterprise to focus on what is important -- which is to provision workloads as fast and easy as you can provision in an Amazon cloud. This levels the playing field with these massive public cloud providers because it empowers IT to offer a similar service with a similar user experience.

Nebula-like technologies also abstract the complexity and cost of standing up these private clouds especially from a security perspective and a maintenance perspective, leaving enterprise IT focused on providing timely services to the business. VMware's entry into this with Project Marvin will only serve to further Nebula-like companies' use case of a modular approach to building on-premises private clouds.

Hybrid angle
Based on my conversation with Kemp, I have a couple of thoughts. First, no conversation about cloud would be complete without touching on hybrid, which has become the new Holy Grail for enterprises. Nebula should help IT teams extend their private clouds to many leading public cloud providers, including Rackspace and other OpenStack supporters.

Now, it's a stretch to call any cloud project "turnkey," but Nebula, from a private cloud perspective, is as close as it gets. Lock-in is certainly a consideration, but realistically, all software locks you in to some extent. Using VMware locks us in, yet it has vast enterprise market share. You can always escape if you're willing to invest the time and money. Same for the cloud.

Last, while technologies like Nebula will accelerate the pace of private cloud build-out, more importantly they force the enterprise to focus on what's important -- making sure the business can provision workloads as quickly and easily in-house as in an Amazon's cloud. This will go a long way to level the playing field between enterprise IT and massive public cloud providers because it empowers IT to offer a similar service with a similar user experience.

GE and Airbnb take very different approaches to measuring cloud computing's value. Their strategies, plus our exclusive survey, show that calculating cloud ROI is far from an exact science. Get the new Cloud ROI issue of InformationWeek.