Skip to main content

OpenStack: Will It Prevent a Cloud Mono Culture?

By August 24, 2010Article

An entire ecosystem of management tools, compatible cloud stacks, and a large number of cloud services are being built around Amazon Web Services (AWS). There are some experts who even believe that Amazon’s EC2 and S3 APIs should become an industry standard. Clearly, Amazon is leading the way with tremendous innovations and can legitimately lay claim to have created an entire market for how infrastructure can be made available in a scalable, self-service manner at highly competitive price points. I also hear from enterprise CIOs that Amazon is very responsive to their enterprise needs and is quickly learning how to work with enterprises to provide a broader and deeper set of product offerings.
Despite Amazon’s significant contributions, most enterprises remain wary about adopting a single vendor architecture and API as the industry standard. Customers are not ready to accept the mono cultures of the past several decades. A recurrence of the proprietary stacks and lock-in of the past are seen as a threat to the value of the cloud.
In my conversations with enterprise cloud decision makers, I consistently hear the same message: customers want choices and options that can best serve their business needs. They don’t want to be offered the proverbial 800-pound guerilla’s menu of, “You can pick any color as long as it’s black.” Without a doubt, customers perceive that they will benefit if there is competition that encourages innovation and dynamism in the market, particularly at this early stage of cloud adoption.
Certainly, mainstream adoption of cloud computing will still take years. Today, many companies use what’s out there simply because it provides massive utility and solves their immediate business problem in a cost-effective way. Current discussions on cloud interoperability standards are really theoretical in nature and revolve around a problem that will only emerge when cloud computing truly becomes mainstream. In the meantime, cloud vendors should continue to innovate and keep things very transparent – these two factors will be key to driving adoption. As adoption increases, open standards will naturally emerge. While premature standards can stifle market innovation, CIOs believe proprietary standards can be worse (and history has proved that the half-life of such standards tends to be very short).
CloudScaling’s CEO on OpenStack OpenStack is an open source, open standards cloud initiative by Rackspace and NASA. To Randy Bias, CEO of CloudScaling, OpenStack is important because it is an opportunity to drive wider adoption of a single, open standards-based cloud stack with an opinionated architecture which will encourage alignment amongst a number of service providers and enterprises in terms of APIs, delivery mechanisms, architectures, and hypervisors.
Can OpenStack pose a serious challenge to the incumbents? Only time will tell, but many industry observers believe that if it can rally an enthusiastic community of developers, customers, providers and standards bodies, and can gather an ecosystem of tools and services, OpenStack has a fighting chance of presenting a compelling option for enterprises and service providers.
“People don’t realize that lock-in actually occurs at the architectural level, not at the API- or hypervisor-level,” says Bias. “OpenStack provides a great opportunity to be an open alternative to the proprietary stacks of Google, VMware, Amazon and Microsoft.”
Large enterprises are adopting cloud technologies in their datacenter in a piecemeal fashion and thereby not gaining the full cost and agility benefits of the private cloud. Furthermore, they have been slow to adopt the public cloud because of the lock-in, security and privacy issues.
Because of these issues, datacenters continue to be as grossly underutilized as ever before. Eighty percent of IT budgets still goes to “keep the lights on” maintenance activities – activities that do not enhance core competency, do not increase competitive advantage, and are not necessary for business fundamentals.
“That’s bad business,” says CloudScaling’s Bias. “IT shouldn’t be spending ten-times more [than what they would spend in the cloud] on traditional systems. Mission-critical systems should stay on-site and everything else should go to the public cloud.” Bias adds, “To get the full benefit of the cloud in the datacenter, IT should not stop with on-demand virtualization; they have to use the techniques that Amazon, Yahoo, and Google use to achieve their price points: Open source software running on cheap, commodity hardware and software engineering techniques designed for failure at scale.”
If OpenStack can rally a critical mass of adopters, and a community and ecosystem that will contribute to it at the same levels as Amazon, VMware, Google, or Microsoft, then we may see the emergence of a true open source alternative that can drive greater interoperability and public cloud adoption through a set of reference architectures.
The Value of OpenStack Randy Bias and Cloudscaling are attempting to leverage OpenStack to help companies build public and private clouds at scale and effectively compete against Amazon, Google, and others. Bias says that OpenStack will allow cloud service providers to quickly build out their non-differentiated offerings so that these vendors can focus on their differentiation in terms of support, service and other value-added service as soon as possible.
There are two components of OpenStack. One is the storage component which is object storage similar to Amazon S3 or Google Storage. This solution comes from the Rackspace Cloud Files project which is actual production code that Rackspace has been using for their commercial offering. Up until Rackspace made this available as part of OpenStack, there was no solution available for companies to roll out their own cloud storage solutions at scale. Out of the box, OpenStack is compatible with Rackspace’s Cloud Files storage.
The second piece is compute, which is used for managing virtualization. OpenStack Compute is a scalable compute-provisioning engine based on the NASA Nebula cloud technology and Rackspace Cloud Servers offering called, “OpenStack Compute”. OpenStack Compute is also compatible with Amazon EC2 in terms of APIs and architecture.
A Look Ahead It’s still early for OpenStack. Features like metering and multi-tenancy will be added later this year. Developers will need to plug in other things like monitoring, image management, OSS/BSS support systems, identity management, authorization, authentication and other components to operationalize the system and to make it production-worthy in the datacenter.
“OpenStack is the third camp – particularly for service providers – that is competitive to Amazon and Google,” says Bias. “Right now, the enterprise adoption of the public cloud is very slow and it will continue that way because most enterprises will go the route of private clouds (either build internally within their firewall or use cloud hosting services from service providers) in the next five years. This gives service providers – including telcos – an opportunity to build differentiated service offerings using OpenStack as a foundation instead of building yet another cloud stack based on their own unique and non-compatible technologies or hypervisors. We think that by using the OpenStack platform, [service providers] can get to market faster with an open solution. This will grow the cloud market faster.”
If you are looking to build a public, private or hybrid cloud environment based on open source technologies, you need to consider your team’s expertise and how much customization you need to input to make it an enterprise-grade cloud. OpenStack should certainly be on your list to experiment with. If your team has Ruby programming skills, Open Nebula is another good option to consider.
SandHill.com thanks Randy Bias of CloudScaling for his insight on OpenStack and the state of the cloud computing market today. Be sure to visit CloudScaling’s popular blog at http://cloudscaling.com/blog.
Kamesh Pemmaraju heads cloud research at Sand Hill Group and he helps companies—enterprises and technology vendors—accelerate their transition to the cloud. He welcomes your comments, opinions, and questions. Drop in a line to kamesh@sandhill.com.