Archive for August, 2010

When did we start thinking that anything less that 100% utilisation was okay?


This will be my last Weekly View for a fortnight as it is that time of year again …holiday!

My fellow Practice Leaders, Paul Casey and Kevin Ebbs, will be enlightening, educating, and entertaining you …all in equal measure, no doubt … with the next few Weekly Views as I recharge my batteries with Mrs. PL and PL Junior.

Before I sign off an help finish the packing, there were a few interesting things that have happened over the past few weeks that I thought I would share with you.

The first is that we’ve launched our Optimised Environment VDC application on the iTunes store as a free of charge download.  I won’t go into this too much as I’ve spoken about this app a few times and even made a demo video, but I did want to thank Nikki Clift and Anna Rosenberg for their help and perseverance to make this app a reality.  It isn’t designed to win any technology or business ‘arguments’ but, rather, serve as a tool whereby one can easily navigate the great collateral we have for Optimising Environments, Virtualised Datacentre, and the like … and I, for one, believe this helps to set us distinctly apart from our competition in the services & solutions provider space as we’re articulating that it isn’t our IP that makes us unique but our ability to execute efficiently and delight our customers with market leading solutions such as VDC and cost benefit underwriting.

The second was Mrs. PL and I attended a ceremony which celebrates an engagement called a vort.  Now, this is a ceremony that has been going on for hundreds of years …possibly longer …but is just as relevant today as it was when it was first thought up and is one of my favourites.  There are many interpretations and much symbolism in the ceremony, but put simply you are firstly giving your word to marry …vort literally means ‘word’ in Yiddish …but it also serves to remind both of the MiLs [mum in laws] that they must be careful with the words they use when their son/daughter comes to seek counsel/complain about their significant other after they are married.  How is this message achieved?  During the vort a plate is broken with each of the MiLs holding opposite ends of the plate …signifying that, once broken, it is impossible to put the plate back together again fully.  And that each MiL has a responsibility for the structural integrity of the  plate and for the counsel they give.

Yes, you can glue it back together …but there will always be the little shards you couldn’t find and it will never look quite the same.  And who wants to eat dinner of a plate glued back together, anyway?!  And so it is with marriage.

What’s this got to with Data Storage & Protection?

I’ve been a geek for as possibly as long as I’ve been alive …Mrs. PL jokes that I had corduroy patches on my pyjamas …and I recall going with my father to work on the weekends where he retensioned the reel to reel backup tapes whilst I played Star Trek on the mainframe.  I was about seven.  Flash forward to my first IT project …a three tier client server project to expand an order entry system from a closed mainframe to an ‘open’ system with CICS, OS/2 Warp 4 servers, and Win95 workstations.

Told you I was a geek, not that I thought any of you doubted me.  But I’ve also spent a fair amount of my life around mainframes.

I’ve always had a soft spot for mainframes and never thought they really got the kudos they deserved, but as with anything in technology …if you wait long enough, the pendulum will swing in the opposite direction and you’ll see the same ideas come back again to solve business issues.

As we democratised technology away from the mainframe and into ‘open’ systems, our job as technologists was to make the pot and allow others to determine how to fill it.  Would Facebook, Twitter, Google, or Wikipedia have happened if the inventors had to submit the jobs on punch cards and provide business justification for the output?  But as data volumes have increased and corporate IT departments have had to scale to meet demand we’ve begun to see how much waste there can be in ‘open’ and distributed technology infrastructures.

Just as we would find it difficult to glue the plate back together if we broke it, so too have we found that it has become costly at scale once we broke the mainframe apart into separate network, storage, computing, and software components.

Now, this isn’t a ‘let’s go back to the mainframe’ rant.  There are many technical advantages to open and distributed infrastructures, but what we’re seeing now with the virtualised datacentre is a return back to a mainframe like device capable of supporting tens of thousands of users from a very dense and efficient footprint.

When did we start thinking that anything less than 100% utilisation on a platform was okay?

I suppose our having had PCs in our homes and on our desktops for almost the past twenty or so years has conditioned us to believe that if we approach 100% CPU, disk, or memory utilisation …bad things happen.  PC slowdowns, mouse won’t move, you know the score.  But mainframes typically operate at 100% utilisation …thats what they are designed to do!  Run multiple workloads on the same infrastructure efficiently.  Yet we know that the average server utilises less than 15% of the total CPU, storage infrastructures operate at best 40% utilisation …put simply, VDC isn’t necessarily a new idea but, rather, we are revisiting the mainframe by virtualising the servers, network, storage simultaneously such that we can run an infrastructure to support 60,000 or more users with just five 40U racks of equipment versus the better than five or more rows of equipment this would take in a traditional distributed infrastructure.

Equally, whilst I recognise that there are some that view VDCs as being ‘cutting edge’ technology they are by no means bleeding edge …indeed, mainframes have been containerising workloads and operating upwards of 100% utilisation for the past forty five years. And the mainframe space is about to get even more interesting as you can run Linux and related workloads on an IBM z196 mainframe.

But I don’t want to steal Martin Boake’s thunder, so if you’ve questions about that or are interested in a mainframe for workloads contact Martin.  He and the team we acquired from Thesaurus are a great bunch of guys and have probably forgotten more about mainframes than most people …myself included …will ever know.

Right …I think that does it for right now and remember, the next time you pick up a plate or consider breaking a workload out …there’s probably a much more efficient way to do that without you having to go and pick up the pieces!

Have a great weekend and see you in a few weeks.

Click here to contact me.