I’ve tried to get my head around all these different buzzwords that we’ve been subjected to for some time and I finally think I have the answer… I could be wrong, but they’re all strangely related and mean pretty much the same thing in many respects, so I’ve decided I’m going to try and define them in this article…
Back in the simple days of computing, we had a server it did one job, had one operating system and had its own disks. Something went wrong, you could pretty much put your finger on it and get the offending part swapped out or the issue with the operating system or application fixed.
Then people started talking about virtualisation, which separated the physical hardware from the operating system by a layer in the middle called the hypervisor. This meant that you could run a whole datacentre that had different makes and brands and should something break you can move the virtual machine onto new hardware whilst you fixed it with little or no downtime.
This of course led storage to be an issue, as if your server running your virtual machine storage systems broke you had a single point of failure, so the storage was decentralised and put into dedicated storage boxes or Storage Area Networks (SANs). Multiple servers could connect to the SAN and pickup whatever virtual machine they were tasked to run, this increased the density and efficiency of the datacentre and people have been talking about this for several years to reduce complexity, increase efficiency and reduce cost. It has also been a big play in disaster recovery and business continuity planning.
Of course the issue with separating servers and storage is that you have to connect them up with something, in the big bad world of enterprise, people used FibreChannel, an expensive networking technology that goes well beyond Ethernet in both price and performance. As time passed then Ethernet also got faster and 10GbE is now available as well as multiple connections and seriously fast switches, which makes the midmarket space for this type of storage set-up much more appealing. Everything got to a nice space of equilibrium and products like Dell Equalogic, Netapp, Overland, Lefthand etc sit quite happily in this space.
However, not to be outdone, there was another jump in the enterprise world of ‘big data’ I’m still confused on that, big datacentres? Who knows! Anyway, what the big boys then started to do is converge switching, storage and computing power via ‘fabric’ and hence fabric computing was born. Fabric computing is essentially a suite of various components that all work together (seamlessly, apparently) so the switch talks to the SAN, which talks to the server, which talks to the virtualisation hypervisor which is all managed from one central console and when something changes all devices are reconfigured and working. It’s associated with big data in that in a large scale datacentre it makes manageability easy.
Not to be outdone, the smaller guys have come up with their equivalent of fabric computing, using cheaper alternative technologies and being relatively agnostic on the switches they use due to the increasing availability of switches with open configuration standards on them. This has been dubbed hyperconvergence – usually done with less boxes and storage and computing in one box, eventually it’s likely some further consolidation will come. Scale Computing, Pivot3, Nutanix and Simplivity are all new players in this space. The main similarity to fabric computing is again the single point of management for everything.
So, we’ve talked storage, fabric, virtualisation and hyperconvergence, but what about cloud? Well, all of these technologies can be considered to be part of a cloud solution, they give flexibility in order create ad-hoc computing and storage resources by an easy to use interface. Most people would refer to this type of cloud however as the private cloud.
So in summary:
- Virtualisation – the consolidation of multiple server operating systems on a single hardware server
- Storage – the consolidation of a business’s storage into a single or multiple storage area networks (SANs)
- Fabric – the interconnection of the above storage and server facilities with high performance switching that is all controlled form a central management console
- Hyperconvergence – much the same, but storage and server facilities tend to be in the same physical box
- Cloud – well anything that has easily configurable and available compute and storage facilities, in business terms this is more often considered to be the private cloud
If you’d like to know more about how the changes in computing technologies can benefit your business in terms of IT agility, cost reduction and disaster recovery we’d love to hear from you at the contact details below.