There has been a tremendous amount of innovation in all aspects of data center infrastructure that addresses the changing needs of the data center. Hardware and software companies have grown to multibillion-dollar enterprises in networking and storage. The only exception has been in the server market. There has only been one company in recent history that has introduced a significant innovation for data centers, the foundational piece of hardware that delivers the value of an application. Yes, new servers have been introduced over the years, but they are essentially the same server brought to market in a different package. Hence data centers are still architected the same way, though speeds and feeds have increased.
AMD is the only company with groundbreaking technology specifically designed to address the next generation of data center computing. It has re-imagined the server with real engineering innovation for Internet scale computing. The SeaMicro SM15000 server with its patented Freedom™ Fabric is the most significant innovation in server technology in recent history, and provides applications, such as OpenStack, the compute, storage and networking flexibility in an integrated package of 64 servers and 1.28 Tbps of throughput in 10 rack units. This technology enables data centers to be optimized for compute (Nova), storage (Swift) or networking (Quantum). Whether it is bare metal (Ironic, a.k.a. OpenStack on OpenStack or OOO) or virtualization, the SeaMicro SM15000 provides scale out capabilities using off-the-shelf components that reduce operations expense by up to 50 percent. No other server can do this.
OpenStack’s future is far from certain, but what is certain is the growth of the Internet and the continuous innovations for both consumer and enterprise services. The key to success is whether the OpenStack community can coalesce and create a foundational set of applications so that companies do not need to create departments for its deployment and management. The current version, Havana, is a big step forward, and the goal posts are visible on the horizon. Deploying OpenStack on AMD’s SeaMicro SM15000 will only accelerate a company’s efforts to reduce licensing costs and realize the benefits of an open source solution. The unique fabric-based design, dense form factor and leading energy efficiency makes the SM15000 server the industry’s best choice for a successful OpenStack deployment.
A lot has happened in infrastructure technology since 90s. We have seen the growth of dedicated network and storage industries. We have seen the rise of server virtualization and the commoditization of servers. We are now witnessing the rise of cloud computing, which effectively commoditizes the whole of IT infrastructure.
But given the broad adoption of VMware in the enterprise as well as the broad adoption of AWS services, why OpenStack? Why all the buzz and excitement?
The answer to that question always boils down to one answer: Cost and freedom! In fact, as Clayton Christiansen has articulated well in his book, The Innovator’s Dilemma, market disruption is always driven by the low-cost technology that becomes “good enough”. Eventually, what was the low-cost technology becomes the market leader.
In today’s cloud world, the thought of having a free cloud platform on which to host one’s applications is equally appealing, when compared to expensive on-premise alternatives, such as VMware, or public cloud platforms, such as AWS. However, the fact that OpenStack lends itself to integration with the likes of Puppet and Chef for workflow automation, means that the potential business implications of OpenStack with respect to application deployment and operation are much greater than what Linux ever promised.
AWS started out by providing a low-cost Platform as a Service (PaaS) environment for developers. Actually, at the time, it was Infrastructure as a Service (Iaas) for developers, now known as PaaS (see more about Amazon’s EC2 history). It has since grown into today’s public cloud juggernaut (see my last blog, in which I gave my take on AWS re:Invent 2014). Its ease of use and pace of innovation make AWS quite compelling. While it is the public cloud market leader, it is not the cost leader. In fact, if one were to create the “killer app” on AWS with tens or hundreds of millions of subscribers and petabytes of content pull every day, it could be quite expensive. Put yourself in Netflix’s shoes: imagine being “all in” with respect to AWS, only to witness the other side of Amazon deploy a competing service (Amazon Prime), running on the same infrastructure. Yikes! What would you do?
Some of that thinking is undoubtedly behind the appeal of OpenStack. It represents not just lower cost, but freedom from vendor lock-in. It allows some customers to leverage AWS capabilities, but gives them the option to pull the production application in-house and run it on OpenStack, if desired, giving the option for the best of both worlds.
As of the Juno release, Openstack seeing great support in the vendor community, providing support for a wealth of server, storage and networking technologies. The addition of Ceilometer for metrics and Puppet/Chef integration for workflow automation strengthen the appeal. We are also witnessing the growth of OpenStack packaging and support companies, such as Mirantis, which make the deployment of OpenStack a 60-minute exercise.
I, personally, am a strong advocate of OpenStack. Here at Ostrato, our goal is to support easy-to-use, strong hybridcloud management and governance, wherever those services live: in public clouds, such as AWS, Azure, RackSpace, SoftLayer as well as in the enterprise, such as VMware. And, yes, we also help govern and manage OpenStack. Check out cloudSM when you get a chance and find out why OpenStack and Ostrato make a great combination.