Shifting to Software-Defined
Showing results for 
Search instead for 
Did you mean: 

How will the next class of infrastructure operate in the Data Center?

Chris Purcell

Tech day 1.png


Last week I had an opportunity to meet with a diverse group of industry influencers to discuss how a new class of infrastructure should operate in the data center…a new class of infrastructure enabled to handle both Traditional Business and the needs of the Idea Economy. This session was different from the focused product and solution briefings I have hosted in the past around Servers, Converged Systems and Cloud. In this meeting I took the opportunity to spend more time discussing different perspectives and strategies around how infrastructure must evolve to accommodate the very different needs IT is now being tasked to support. The discussion centered around Composable Infrastructure which Hewlett Packard announced in June of this year at its big technology event in Las Vegas.



Tim Crawford  @tcrawford                       Rich Miller  @rhm2k

Mark Thiele  @mthiele10                          JP Morgenthal @jpmorgenthal

Ed Horley  @ehorley                                Rob Hirschfeld  @zehicle

Philip Sellers  @pbsellers                          Roger Singh  @rogerwsingh

Jake Ludington  @JakeLudington              Cameron Cosgrove  @cameroncosgrove

Dion Hinchcliffe @dhinchcliffe



Why discuss Composable Infrastructure?

Contemplate for a moment the explosive growth of the Idea Economy, where we expect to see 7.6 billion people, with 100 billion connected devices and things…requiring 1 trillion applications by 2020. Now think of IT as it is today and the expectation that it will provide the infrastructure necessary to support deployments of new mobile and cloud-native products and services while continuing to maintain existing traditional IT applications. It’s somewhat surreal, isn’t it? But that’s where we’re headed. IT will have to split resources to support two different modes of infrastructure – one mode to support traditional applications where change is infrequent, and reliability and stability are primary goals, and the second mode where supporting new apps and services is the norm and flexibility is the key requirement. The changes being forced by the Idea Economy and the ability of IT and business in general to handle them made for a very interesting and lively discussion.


I put forward that Hewlett Packard Enterprise’s vision of Composable Infrastructure makes it a very viable solution for bridging these two modes of operation. The concept is simple; physical and virtual resources can be provisioned to applications on demand regardless of the style of application or the type and amount of resource needed. Hewlett Packard Enterprise defines Composable Infrastructure as providing fluid pools of compute, storage and fabric resources that can be aggregated and disaggregated on demand, abstracted through a unified API, and controlled through the use of intelligent automation.


Composable Infrastructure briefing

After doing a brief round of introductions with the group, the conversation started to go full force around the need for a new type of infrastructure – why the industry needs it and how it should operate. As you would expect, there were many differing opinions from just a software defined layer, to taking a hybrid approach, to some saying it would probably be too hard to move the data center onto a new platform and create a new standard to operate within.


From there, we continued to unpack the topic and talk through the three differentiators of Composability and its basic principles – fluid resource pool, unified API and software-defined intelligence. It’s the combination of these three principles working together that defines composability. Anything less may provide some degrees of Composability, but will not provide the full benefits of hybrid/bi-modal IT from a single Composable Infrastructure, which is what is needed by IT.


Fluid resource pools are fluid pools of compute, storage and fabric resource that can be continually aggregated, disaggregated, and re-aggregated to support the needs of specific applications and services

A single unified API is a simple and open, industry-standard REST API that provides a single line of code to abstract every element of infrastructure, enabling fast, smooth integration and automation of resources

Software-defined intelligence enables template-driven, single-touch provisioning and management that provide true infrastructure-as-code capabilities


Comments from the briefing

“The real differentiating value over time for the industry will be from a truly shared common platform. One that can wield -- and offer up as an unbeatable differentiator.” Dion Hinchcliffe @dhinchcliffe


“Software is the big differentiator going forward. To some degree I subscribe to this level of thinking as well. If HPE can find ways to make their HW platforms participate in a software-managed scalable resource pool, that's great, I just don't believe it's a selling feature. The management layer that links into to manage the pool is the thing that will be the key differentiator.” JP Morgenthal @jpmorgenthal


“I'm looking forward to adopting HPE OneView, templates being the second.  These are the pain points operations deals with.  They're tangible improvements that the software can bring to market.  I'd encourage HPE to look at more of these kinds of improvements to appeal to ops teams.” Phil Sellers @pbsellers


“The concept of a ‘floating think-tank’ approach to introducing a product or product strategy was very good.  Also, by establishing at the outset that nothing was ‘sacred’ and everything was open to question and inspection was refreshing.” Rich Miller @rhm2k


“By using a single infrastructure stack for all types of applications, customers are no longer worried about stranded resources as applications move from legacy to newer architectures. Resources are immediately available for repurpose via the resource pools.” Tim Crawford @tcrawford


“While a lot of practical questions would still need to be sorted out, HP's concept is fundamentally on the right track. Composable Infrastructure is disruptive because it provides capabilities that are available in the cloud, but not on premise. "So this is a mindset switch." Cameron Cosgrove @cameroncosgrove



The briefing went much better than I expected. I felt that everyone had a lot of respect for each other and at times when the conversation got very intense, a lot of great ideas came forward. I liked Rich Miller's @rhm2k mental image of this group as being a ‘floating think-tank’…as that is how the meeting ultimately unfolded. It allowed me to test current messaging and receive unfiltered feedback on how HPE is currently constructing the story around Composable Infrastructure.


The other exciting aspect of this discussion for me was that it’s not often you get an opportunity to be involved in the birth of a new class of infrastructure, one that will reshape the data center as we know it today…and I think the group felt the same way. 


As HPE continues to roll out more Composable Infrastructure, customers will continue to see that it has the unique ability to run without being limited to one computing paradigm because it can run virtual machines, bare-metal deployment, containers, and cloud-native applications. It can run anything and store anything. It also fully integrates with industry-standard tools like Chef, Puppet, Docker, VMWare, OpenStack and others, allowing IT operators and developers to continue leveraging the tools they use every day.


Clearly there are some very exciting times ahead of us…and again, thanks to the group that took the time to participate with HPE in Palo Alto. It was a great day.




Keep up with HP Converged Systems on Facebook

Follow us on Twitter @HPConvergedSys

Join the Converged Infrastructure discussions on LinkedIn

Learn more about HP training classes at HP ExpertOne

About the Author

Chris Purcell

Composable Infrastructure, Integrated and Multi-Cloud management, Hyperconverged Infrastructure and Cloud