WebScale & the SAN Hangover

A three node cluster

A three node cluster

Before I joined Nutanix one month ago I was trying to explain the concept of WebScale to a friend and as with most things I used the analogy of going for a beer.  It’s currently 27 degrees in the UK with nothing but blue sky and there are few things better to help along an English summer along than a cold pint.

The traditional way of buying storage + compute is like ordering ten pints of very strong ale all at once because you might possibly, over the next several hours of an evening, drink that much.  And, as with most IT SAN purchases, you may as well buy the strongest and most expensive one on the menu because you don’t want to be left thirsty – or out of grunt and capacity.  The reality is most organisations either don’t know their final requirements – because things change rapidly in this world – and run out of resources or more likely they never actually use it all and end up wasting their money.

In short, SANs are like a table of warm stale beer.

There’s also so much complexity in the old ways as well with fabrics, bespoke hardware, LUNs, RAID groups,

A three node cluster with a node or CVM failure.  Oh look the data is safe and available!

No loss of data

aggregates, zoning, configuring failover, presenting it all to the hypervisors and actually getting your SAN team to talk to the virtualisation guys.  Back at the bar you’ll have to learn how to pull a pint, build the brewery, take up farming to grow the hops and maybe even employ another farmer and barman for when you go on holiday or fall ill.  Worse still if you don’t like the taste after a year then you have to do it all over again because upgrading means a disruptive replacement process.

The Nutanix approach to scaleability – one part of what we define as WebScale – is paying for one drink at a time.  If you feel thirsty after the first pint – you need more compute and storage – you just order another node, plug it in, do a couple of clicks in Prism and you’re done.  Another cold pint of Nutanix goodness with minimal effort and a Nutanix cluster expanded in moments with zero down time.

A three node cluster expanded  to four, with no downtime.   Wonderful!

A three node cluster expanded
to four, with no downtime.

If you decided to change from beer (NX3000) to Redbull (NX1000) to wine (NX6000) to Jager (the NX7000) Nutanix also allows you to mix your drinks…without the hangover of having to re-do all that work.  Come to think of it a few customers have mixed the NX1000 with a couple of NX7000s for GPU acceleration to make Nutanix Jagerbombs 🙂

Managing a SAN, especially speccing it and scaling it to last over several years as a business or project grows, seems like a lot of hard work to me.  Unnecessarily so.  The Nutanix platform is predictable and scalable and that’s just one thing that Webscale means to me and is enjoyed by our customers and partners.




1 Comment

  1. Gabi


    Nice post !

    I feel I can really relate to this, through your alcohol analogy.

    Good read, hope the new job is going well so far.



Leave a Reply

Your email address will not be published. Required fields are marked *

© 2021 Nutanix Noob

Theme by Anders NorenUp ↑