OpenStack For Private Clouds
On February 15th Tom Bittman of Gartner published a blog which asserted that 95% of Private Clouds are Failing. When an industry analyst makes a statement that big, in one of the top three priorities for enterprise CIOs today, it’s critical that we as an industry step back and understand how we are going to help our customers move from a 95% failure rate to a metric of 95% success rate in private cloud deployments. In my role at Cisco, it is critical that I understand some of the challenges customers face in their cloud journey and help them navigate through the complexities towards success. I’d like to share some of the things that I see happening in the industry that are helping to solve the exact problems Tom points out in his blog.
Let me explain:
Focusing on the wrong benefits: Public or private cloud should be about agility and Fast IT—not OPEX vs. CAPEX. Chasing a penny is never a good reason to implement a technology and, in many cases, it ends up costing you more in the long run. Case in point, we saw this in the Linux days where people chasing “free” were often left in a hole while those chasing “freedom” ended up succeeding in their endeavors. Cloud is in that exact same phase today. Customers chasing low cost CAPEX today can find themselves locked into a fixed architecture that ultimately costs them three times the cost of running it themselves. First and foremost, a cloud operating model is about providing agility to your teams to innovate quickly and without the incumbencies of a classic IT structure. As I mentioned in my recent blog on hybrid cloud, this is what drives people to the public cloud in the first place. But sometimes a public option isn’t viable (compliance, security, control, etc.). The latest trend in helping drive successful private cloud deployments in the enterprise goes well beyond virtualization—it follows a managed private cloud solution that brings the true cloud model in-house and enables development teams to get all the agile IT benefits of the public cloud (automation, self-service, on demand) within the confines of their own datacenter. Not only are we seeing more successful deployments, but customers that are chasing the right benefits—including agile IT behind the firewall (freedom)—ended up reaping the cost benefits of being cheaper than a public cloud option.
It’s not really a cloud: I’ve seen many cases where IT vendors mischaracterize a highly virtualized environment as a private cloud and leave their customers to suffer the consequences. Virtualization is a core foundation in cloud computing, but just because you are virtualized does not mean that you are a cloud. No matter how much the virtualization market has evolved, it still does not have features such as self-service or rapid elasticity, which are key tenets of a true cloud. Enterprises looking to bring the power of the public cloud behind their own firewall need to be partnering with IT vendors that are giving them a real cloud architecture vs. more intelligent automation. As a starting point, there are many vendors today building out enterprise cloud solutions based on the OpenStack Project, the largest open source cloud computing project in the world. Not only was this project born in the cloud era, it embodies all of the core tenants of cloud including self-service, on-demand, elastic and agility that enable an internal IT group to bring the power of a real cloud to the user in their enterprise, but with the controls around costs, compliance and security that they absolutely need to be able to guarantee back to their organization. For enterprises looking to take advantage of the groundswell behind this exciting project, there are many great companies to choose from. While I have my personal favorite with the Cisco OpenStack Private Cloud offering (formerly Metacloud), companies like Blue Box Cloud, Mirantis and Red Hat have also been demonstrating successful cloud deployments in the enterprise by taking their experts in cloud computing and applying them to the OpenStack project. The end result is that customers get an open and standards-based real cloud solution that they can successfully run within their own data centers.
Operating a cloud is difficult: An enterprise cloud project requires infrastructure and operations teams to work together in order to ensure a successful deployment within the enterprise. Bittman articulates it well in his blog, “People are your biggest supporters, or your biggest roadblocks.” The challenge most enterprise IT managers have is that they have never run a cloud before. In this market transition, this is one of the biggest skill gaps in the enterprise and also one of the most important ones to ensure success in a cloud deployment. A fairly new approach in the market is around a managed private cloud – infrastructure that sits behind a customer’s firewall but is managed remotely by a team of cloud experts. This is where we are seeing the biggest success with companies adopting cloud because you don’t need to be a cloud expert to have a successful private cloud. The managed cloud model allows a vendor to partner with customers and bring the cloud experts to them—in a cloud like model. Think of it more as “Cloud Operations as a Service,” on-demand and immediately available to the customer. Not only do customers get the “real cloud” environment, but they get a team of cloud experts to remotely manage and operate their private clouds complete with SLAs and support. This is one of the primary drivers behind the success in our customers’ private cloud deployments.
It’s not just about cloud—it’s about an IT strategy: My favorite line in the Gartner blog is: “Optimizing for everything means optimizing for nothing.” Customers have to find the right balance between an agile compute model and classic IT process, as cloud is not a cure-all for enterprise IT. Peder Ulander, who works on my team, once wrote an article on cloud computing for CIO Magazine where he declared that enterprises don’t need a cloud computing strategy. While he wrote this article four years ago, the advice he gives is still sound today: “rather than looking at the cloud as a separate replacement strategy, companies need to look at it from the bigger picture as a complete IT strategy.”
The success of private cloud deployments is really important to us at Cisco. Our strategy is to provide modern development teams with the tools and solutions that help them move faster – enabling everything from hybrid cloud computing to Fast IT and into the future with Internet of Everything. In order execute on this strategy, we have to help solve the “95% failure” problem that Tom Bittman articulated in his pose. It starts with helping customers understand these best practices for bringing the power of the public cloud into the control of IT. If you’d like to read more about the nuance of creating a private cloud that delivers a true public cloud kind of experience in a customer’s data center, Scott Sanchez just published a blog on this topic “Why Compromise Your Public Cloud Experience?”
By Nick Earle
Nick Earle leads Cisco’s global Cloud and Managed Services Sales and Go To Market (GTM) strategy.
Previously at Cisco, Earle led the Worldwide Services Field organization, during which time its revenue grew to more than $10 billion. He also has led the vision to deliver new business models, solutions, and strategic partnerships in the Global Enterprise Theatre. Prior to that, he was responsible for the European Services Sales business in 20 countries. Common to all roles, he and his teams have continuously helped customers solve their toughest business challenges, and enabled our partner community to improve their profitability, using Cisco technologies and solutions.
Prior to joining Cisco in 2004, he served as CEO at StreamServe Inc. and as President of EMEA Operations at Ariba, as well as CMO for Hewlett-Packard’s $35 billion enterprise computing business. He is also a recognized authority on emerging business models and has co-authored two books: Mesh Collaboration: Creating New Business Value in the Network of Everything (2008) and From Dot.com to Dot.profit (2000).
Earle holds a first class honors degree in computing and an honorary doctorate in computing from Liverpool University in the United Kingdom.