Hyper Converged Now and Next – Troy Mangum – Ep49

The IT industry is full of new trends, some you get, some you don’t, one such trend, that until recently I didn’t really get, was Hyper-Converged, a new market, with a message of simplification and dominated initially by new technology players, like Nutanix and Simplivity (now part of HPE) and they have been pretty successful, so why have I not gotten onboard?

A good test with any new technology is does it solve a problem or improve the way I currently do things? Up to now with Hyper-Converged Infrastructure (HCI) I’m not sure it really does, is it helping me build a more automated, flexible, agile IT infrastructure? Is it helping me build a hybrid environment? Is it automating my IT environment so that my business gets the agility it wants? Not sure.

What HCI does do well is simplify your hardware infrastructure, takes something that may have been installed in a full rack and squeezes it down into 2 or 4U in a single chassis, with compute and storage integrated together and a scaling model which allows you to attach another HCI box and scale your compute and storage again.

But is that enough? When I’ve worked with organisations considering HCI, the cost of this model tends to be inline (if not more expensive) with buying the individual components and installing them yourselves and unless those accounts have been looking to refresh compute and storage at the same time, the value has been hard to find.

What’s changed my view? The starting point is nothing to do with changes to the HCI hardware model or addition of some great new feature, it’s actually and maybe not surprisingly driven by software, look at what Microsoft and VMware are doing for example, VMware is delivering an increasingly more software-defined infrastructure with every incremental release of their virtualisation stack.

Microsoft’s Azurestack, although limited currently, aims to bring a fully software-defined Azure like experience onto your local hardware and of course solutions from both of these companies are increasingly hybrid focussed, VMware on AWS and of course Azure both integrated tightly into these on-prem stacks.

This simplification of the software stack is now starting to drive the need for a hardware stack that matches this simplification and can take advantage of these software-defined infrastructure solutions.

It is this changing environment that is the focus of this latest podcast.

At the recent NetApp Insight conference, I met with Troy Mangum who shared some research he’s been working on reviewing the HCI market, how it stands today and the changes HCI  vendors need to make to ensure they build on the early success of first-generation solutions to deliver a platform to meet the needs of the modern data centre and take advantage of these software-defined infrastructure stacks.

We explore a range of discussion points from the research, we look at the drivers behind the adoption of HCI, the need for simplification and easier consumption of IT resources. We also discuss how the current technical design of HCI hardware architectures may limit their ability to grow in the way we need them to.

Troy shares how currently HCI comes with a risk of introducing infrastructure silo’s into our datacentres, focussed on solving individual problems and not the flexibility the modern data centre needs, we also explore the phenomenon of HCI tax, what this is and why it’s a problem.

Finally we take a look at the future, how architectural changes are driving a new breed of HCI architecture, a second generation, allowing a more flexible deployment model, decoupling the component parts so HCI can scale capacity and compute separately, which then begs the question, is this new breed of HCI really HCI at all and does it really matter? And of course, we look at NetApp’s entry into this market.

To find out more on this topic and what NetApp are doing you can find lots of HCI information on NetApp’s website here.

You can also find out more from Troy by following him on Twitter @cloudreveler

Next week we look at very large data repositories, as I’m joined by returning guest Justin Parisi to discuss the concept of Flexgroups.

To ensure you catch that show, you can subscribe to Tech Interviews on iTunes, Stitcher and Soundcloud and all good homes of podcasts.

Thanks for listening.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.