Image of fast moving traffic.

by Clare Bourne

As enthusiastic adopters of next generation technologies here at Redcello – we would like to bring to the table some of the technologies that we are currently looking to bring to our customers. Most recently we have been investigating the use cases and technologies behind HCI (hyper converged infrastructure) and how this technology can help our customers, prospects and our own infrastructure projects. Many MSPs are now adding HCI products to their portfolios and vendors are now bringing these products to the forefront of the infrastructure technology market.

What is HCI?

Hyperconvergence is a testament to the evolution of the software defined data centre. For those of you for which HCI is a new concept, it all kinda started like this: Traditionally, infrastructures in data centres were made up of networking hardware, servers and storage – these 3 core elements were likely to be from different manufacturers and required extensive testing and set up to ensure they were interoperable once in the rack. These components were also likely to have to be managed separately from each other.

Next in the chain of evolution came the converged infrastructure – here a bunch of vendors got together and decided to take pity on engineers and provided the traditional compute, storage and network components - but ensured that this hardware could integrate with ease. Engineers could also enjoy a unified support solution from vendors. Pre-packaged and simplified with a single virtualisation and management layer.

Now this is where it gets juicy – HCI (hyper-converged infrastructure) is an infrastructure system again consisting of the traditional compute, storage and networking – however this time it’s all rolled up into a single modular appliance, all pre-configured with a software defined layer and management system. Plug it in and off you go, great right?!

So, what’s the hype about?

There are obvious benefits that we could wax lyrical about here, let’s get started:

The agility of it - It’s easy to set up and run, you can scale it and manage it from a single pane of glass. The pool of compute, networking and storage are managed by a single hypervisor. The easy scalable “node-ness” of it makes it easy to get more resource quickly and cost effectively. The ability to move workloads is simple.

Improved operational efficiency – Not only will HCI save you money and time, it also purports to increase your productivity too. Software defined data centres boast many automation and orchestration tools that improve productivity as well as cost reduction (run rate expense and opex can be mitigated).

Updates are easy and new features aplenty – Due to the software “definedness” of the hyper-converged infrastructure, it’s much easier to get necessary updates and new features in software releases, which are practically and instantly enjoyable.

It can handle intensive workloads – Due to its elasticity and the ability to optimise its IOPS capacities, the HCI can handle boot ups and login storms with ease. It can also reassign storage capacity and shift compute power around to where it’s needed most.

What’s bad about it?

Reading further between the lines there are some signs here that tell us that our rose-tinted spectacles may be a little cloudy. Let’s face it – it could almost be too good to be true.

Change Management. - Once it’s in it’s probably a dream – however, to get to that point how do we broach the legacy equipment? A rip it and skip it job of course costs time and money. Also, isn’t it much kinder to the world to get the old stuff to its end of life first? Further to this, there are still some challenges in setting up a HCI to get working optimally for your specific workloads.

Do your sums - A HCI system can have performance restrictions and will sometimes not meet the demands of some applications. You will not be able to independently scale those individual components (or change vendors for those) - its also key to understand the overhead on the hosts, what the network requirements will be and calculate the usable disk space, not just raw capacity.

Oh, the nuances – although HCI may at first be a “buy a box, plug it in, off you go” solution there are many, many nuances to address before you get the most out of it. Thinking about the complexities of different workloads and how they run and at what times – which one gets first dibs on resources, what workloads are priorities are just an initial set of questions out of the plethora of nuances you may need to address.

Is it really right for you?

Depends on what you're running on it - The node based scalability makes it great for hosting virtual desktops where you know how many virtual desktops per node you can accommodate. It’s also great for business continuity strategies as some vendors provide pre-configured value added features such as data replication and recovery.

Overhauling your infrastructure – If it’s time to get shot of that legacy end of life hardware for good and your IT strategy is to move forward with your own private or hybrid – this is an easy start to that end.

If you would like to discuss your infrastructure with us here, we love chatting about anything (but IT is our forté). Please feel free to get in touch.

Contact us

We would love to work with you!