Archive for November, 2009

To a worm in horseradish, the whole world is horseradish. Or why IBM XiV is still relevant.

30/11/2009

Mrs. PL and I have been trying to add another PL Junior to our tribe.  That’s the good news.  The bad news is …and how do I say this …we’ve had some very robust conversations as of late regarding upgrading to a larger house to accommodate a new PL Junior.  I think it is commonly referred to as ‘a full and frank discussion’ in diplomatic circles  …all I know is I’ve been getting my not inconsiderably sized posterior whupped regularly in our little fireside chats.  Oxbridge debating teams have nothing on a determined Mrs. PL, in my opinion.  Truth be told, I can kind of see Mrs. PL’s point and, to be fair to her, she is genuinely interested in more space as opposed to playing postcode bingo with the yummy mummy brigade who inhabit our little corner of northwest London.

Whilst we have been married for six years and together for almost ten, I still naively cling to the belief that if I just keep talking and present a coherent and factually based argument that Mrs. PL will come round to my way of thinking.

Me: ‘But we can’t really afford a new house, and I’ve been upgrading our house recently …what about our new supercharged home office?’

Mrs. PL: ‘Nice try, but weren’t you …by your own admission …indulging your own inner geek?  How does you being able to Twitter or tweet or whatever the heck it’s called build a new baby room?’

Me: ‘Yeah, okay …but what about the new shower stall, or the new washing machine?  It spins at 1400 rpm!  And has a 20 minute steam cycle to freshen up shirts when they’re wrinkled!’

Silence.  I’m pretty sure Mrs. PL is melting my inner organs with her glare.

Me: ‘And what about the new refrigerator?  It’s like a magic superfridge made by wizards and Hobbits …nothing ever goes bad in there!  We’ve eaten things that are like three weeks over the use by date!’

Mrs. PL: ‘Tell me something, my dearest chucklehead.  How do these upgrades fit into this equation you keep banging on about?  Wouldn’t a new house as opposed to siloed upgrades have a better five year cost benefit?’

Silence.  I hate it when she’s right.

What has this got to do with data storage and protection?

Mrs. PL has got me thinking about Howard Moskowitz, horizontal segmentation and a great talk I heard from Malcolm Gladwell.  I don’t want to steal any of Malcolm’s thunder or take too long explaining horizontal segmentation so click here if you have about fifteen minutes, well worth your while.

Put simply, the thought is that there is no such thing as a ‘perfect’ product nor, by extension a ‘perfect’ solution.  Rather, each product or solution should be developed and subsequently recommended based upon the good it can do for a particular customer situation.

I was reading a blog post from an analyst recently which questioned Is IBM XiV Still Relevant?  Whilst the blog post makes some interesting points, I kept coming back to the same thought …yes, I suppose you could ask this question but only really if you are viewing IBM XiV next to other storage array products in a ‘bikini contest’ fashion.

But judging arrays in a Miss World style lineup isn’t the real value of grid storage …and not at all the way I would advocate our articulating a solution in any case.

No, I think that grid storage …in this case IBM XiV, but you could also make the same argument with EMC Vmax or NetApp ONTAP v8 …is a basic building block of the virtualised datacentre.

If we wanted to view IBM XiV as a building block, one of the more interesting announcements around IBM XiV was actually buried in an announcement IBM made on 10 November which was talking about asynchronous mirroring …but the very next paragraph of the announcement talks about new support for instant space reclamation.

Why is this important?  Well, if you think back to this post about thin provisioning, what this means is that IBM XiV is making the software APIs which make thin provisioning possible available to third party products such as Symantec Storage Foundation such that Symantec software can now ‘recognise’ unused space and return it to the storage pool quickly.  We could easily add an IBM N Series gateway to provide NFS/CIFS NAS in addition to the block level storage from IBM XiV, as well as Storewize to give us data compression from 45% or higher for stale data.

What would this give us?  What we want …and need …to see, with vendors working together to ensure their products ‘glue’ together such that we can build a horizontally capable virtualised datacentre which is efficient, optimised, and fully flexible for customer needs both now and in the future.

But we wouldn’t stop with just the first building block if we wanted to derive true ROI and cost benefit.  We would need to consider virtualising the servers, converging the network, optimising the physical servers with blades, and automating the whole lot.

And here’s where it could get tricky if we start trying to articulate such a solution with stories of Prego, Howard Moskowitz, or Malcolm Gladwell and horizontal segmentation.

I think one of the easiest ways to visualise this concept is to picture a virtualised datacentre as a solved Rubik’s cube with each of the six sides a different solid colour made up of nine blocks.  Each solved side represents one of the discipline areas required for a virtualised datacentre …Data Storage & Protection, Networks, Platforms, Virtualisation, Automation, and Workspace / Collaboration.

Our customers …all of them …have unsolved Rubik’s cubes with the coloured blocks in any of a number of different iterations.

Our job, in my humble opinion, is not to articulate a storage product …or products …in the context of the proverbial bikini contest but, rather, in the context of exactly how our recommended solution will help our customers solve one, then two, then three sides until they reach all six for a fully virtaulised datacentre which delivers true ROI, cost benefit, and little or no disruption to their production business.

Please contact me if you would like assistance in taking this journey.

-Matthew

Click here to contact me.

Advertisements

Magic Quadrants are good, but working equations are better.

21/11/2009

For those of you who know me well …or have joined me for a Chief Wine Officer event …you’ll know that my two favourite hobbies which I’m most passionate about are aeroplanes and wine.  Not always in that order, and never enjoyed together as the Civil Aviation Authority takes a dim view of such interaction.

But I’ll tell you a secret …I don’t actually have a favourite bottle of wine, nor do I have any silly rules like ‘no bottle under £20’ or some such.  I admit that I do subscribe to Decanter, read Jancis Robinson online, and subscribe to more wine blogs and Twitter feeds than I care to mention.  Whilst data can be very useful, you always run the risk of ‘analysis paralysis’ and, at the end of the day …much of what is written about wine is frankly someone’s subjective opinion.

No, I firmly believe that wine should be had for enjoyment …and I’ve tasted exceptional wine at £5 as well as wine costing much more which I wouldn’t clean Mrs. PL’s motor engine with.  Equally, as each one of us has an idea of the tastes we like and don’t like …who am I to tell someone else whether a bottle is good or not?  All I can do is tell you if I like it, although this does introduce the small problem of what to serve at dinner parties or when Mrs. PL and I are sharing a bottle.

So, what to do?  I do have a little formula in my head that I use which takes things into account when I choose a bottle …why are we drinking this, is it a celebration or a weekday? …how much does it cost, and is that a fair price? …what kind of food are we eating, or are we not eating until later? …and so on.  I want to get on to the crux of this post, but at the bottom of this post there are a few wines which make the PL Wine List.

What has this got to do with data storage and protection?

I’ll tell you another secret …contrary to what some might believe …including a few of our vendors …I don’t have a ‘favourite’ vendor or product any more than I have a favourite bottle of wine.  Without getting too Eddie Haskell about this, what is truly important to me …and I know I am far from alone in this within Computacenter …are our customers and how our solutions can help them remain competitive in their respective markets in the midst of a difficult economy.

Great, fantastic, huzzah.  But so what.  Isn’t that, you know …your job?  Indeed it is, but just as it can be difficult to select a wine for an occasion where it will be shared with others …how do we select a solution for a customer in a selective and demonstrably valuable way?

Some customers work directly with vendors and often use Gartner Magic Quadrants as a way to select their preferred solution.  Nothing wrong with that, but just as some winemakers and wineries are now openly criticising scoring systems they see as subjective scoring techniques such as the Robert Parker 100 Point Scale …so too are some vendors criticising the Gartner Magic Quadrants claiming the research methodologies are something less than scientific.  Indeed, a vendor recently brought a suit against Gartner claiming exactly this, with the suit having been initially thrown out but likely to be appealed.

Now, this post isn’t about criticising or having a go at Gartner or their Magic Quadrants …indeed, I applaud Gartner for being very open and transparent regarding their research techniques leaving folks to make up their minds for themselves.

That said, I believe research provided by companies such as Gartner to be but one part of the solution equation.

In an effort to inject more science into a solution decision, rather, I would argue that the solution equation should be expressed as [ROI] + [CBA] + [DPB] = CSS.

ROI, or Return on Investment.  How does our proposed solution return ROI within our customer’s stated period?  How can we leverage the existing infrastructure and investment to improve upon the ROI period?

CBA, or Cost Benefit Analysis.  Once the solution has been implemented, how much cost can be removed from our customer’s infrastructure and related budgets?  Exactly how will this be achieved (e.g. thin provisioning, data deduplication and/or data compression, storage virtualisation)?  What is the CBA not just for one to three years, but for five years from implementation?

DPB, or Disruption to Production Business.  What disruption is the recommended solution likely to have on the customer’s production business?

We give each of the above blocks …[ROI], [CBA], and [DPB] each a possible score of 100 such that a perfect solution would give us 300 expressed as CSS, or the Composite Solution Score.

How do we score each of the blocks such that we aren’t scoring subjectively?  Well, firstly we ensure that our data consultants retain the highest credentials in the industry …but we then couple their knowledge with a point system derived from IDC Storage v3.0 criteria as well as the Carnegie Mellon Capability Maturity Model.

The findings, CSS, are then presented to the customer in either a ‘leader table’ format or as an executive review comparative matrix based upon the vendor solutions the customer informed us they were most interested in.  In addition, the findings often form the basis on which we can offer to underwrite / gainshare the proposed cost savings for up to and including five years from implementation.

How do we do that, exactly?  Well, I can certainly provide you some samples but it does very much remain Computacenter intellectual property …and it probably doesn’t hurt to have a Practice Leader for Data Storage & Protection who studied neuroscience and Technology Leader for Data Consultants in Bill McGloin who studied applied mathematics.

Does it always work?  Yes …and no.  Just as people have reasons for liking or disliking different wines, so too customers will have reasons for agreeing or disagreeing with our findings.

But I believe this is just about the fairest way I know to present a proposed solution in an agnostic way …and, at the very least, absolutely articulates our value to a customer as a true service and solution provider.

As always, if you would like further assistance in taking this journey please don’t hesitate to contact me.

Have a great weekend,

-Matthew

Click here to contact me.

PL Winter 2009/10 Wine List

If you like champagne as a pre-dinner drink or to celebrate, you won’t go far wrong with Heidseck Monopole Gold Top vintage 2004.  Always a quality drop, but at £19.99 from the normal £40 …or even £18.99 if you are near a Budgens …this is a steal!

If you are looking for something more ‘unique’ in the champers department, why not try Nyetimber?  Produced in Sussex, which has geographical features identical to the champagne region, this the the tipple good enough for Her Majesty to serve at the Royal Garden Parties.

I’m a huge fan of pinot noir from New Zealand, and you won’t go wrong with the bottle which won the International Wine Challenge for Best Red Wine, Wild Earth.  At £18 a bottle …and if you hunt around I’ve seen it as low as £13 …how affordable is the world’s best red wine?!

One of the most complex and interesting reds I’ve recently is Lillian Shiraz Mataro 2005.  At £11.75 a bottle, I challenge you to blind taste it and tell me it doesn’t taste about three times more expensive.  I’m stocking up on this one!

Finally, to round out the reds I give you Château Méaume ‘Château Matured’ 2003 Bordeaux Supérieur.  A bit of a mouthful for a wine costing a very affordable £8.99, yet if you open it 45 minutes before dinner I guarantee your guests will think you spent a whole lot more than that!

The end of history and the last technology?

13/11/2009

Francis Fukuyama wrote The End of History and the Last Man in which he famously postulated that, as far as systems of government and markets were concerned we had reached the ‘end of history’ …democracy and the free market had defeated all comers and, as far as Francis was concerned, the only thing left open to debate was how to implement such systems and what controls were required for regulation.  The book became a bit of a touchstone and rallying point for what would become known as neo-cons, but as time has marched on many of Francis’ original assertions have been challenged by the likes of Robert Kagan in The Return of History and the End of Dreams …as well as real world events like 11 September, the Iraq war, and the recent economic recession.  The end of history?  Doubtful.  Just as surely as we have debated systems of government and markets since Greek and Roman times, so we will surely debate them in the weeks, months, and years ahead. 

What has this got to do with data storage and protection?

There have been many exciting developments in technology generally and storage specifically over the past few weeks.  VMware, Cisco, and EMC …the VCE consortium or coalition, if you will …announced a reference architecture which is essentially a virtualised datacentre in a box called vBlock as a infrastructure package / product in its own right, with Acadia as a private cloud ‘solution provider’ to help enterprise customers migrate their existing infrastructures to said virtualised datacentre and possibly even the much vaunted cloud.  Not to be outdone or left behind, HP also made a similar announcement regarding their converged infrastructure, and then made things very interesting by acquiring 3Com for $2.7 billion.  We also had IBM with their development and test cloud launch and the IBM Cloud Academy, and we mustn’t forget BT having selected NetApp as their provider of choice for BT cloud offerings labelled the Virtual Datacentre.

Now, announcements are all well and good but there remains plenty left for us to more fully understand about just what these solutions will look like …and cost …so I don’t want to get too far ahead of myself nor overstep my bounds prior to more senior Computacenter executive views on what this all means for both us and our customers. 

That said, I am hugely excited about what, at first glance, would appear to be the conversion of hype to reality.  We’ve heard so much about ‘cloud’ computing over the past few years that customers were becoming palpably sceptical …in many cases, cynical …about the ‘cloud’, and I can’t say that I blame them.  Indeed, if you want to have fun ask two industry analysts or vendors to defined cloud and watch the three or more definitions come back.  I say fun, but I’m a geek remember so I tend to find analysts and vendors arguing about these things humorous …but customers don’t and, frankly, don’t really care about the new whizz bang features of a cloud solution.  No, what they care about is how technology can improve their business …or Sharpen it, if you will …to reduce their costs in the pursuit of their being competitive in their respective market.

So are these announcements the signalling of the end of history for technology …will the cloud be the final end point for customer computing infrastructures and datacentres?  I personally don’t think so, and tend to view the announcements as a watershed waypoint rather than an endpoint.  Just as I don’t believe that Spotify will be the death of iPods or iTunes, nor do I believe that customers will take an ‘all or nothing’ approach to cloud computing.

If you look more closely at the recent vendor announcements, what is perhaps most interesting is that the productisation of a virtualised datacentre through the use of a reference architecture is essentially made possible by aligning core products such as VMware, a scale out storage offering, blade servers, a unified network to bring it all together, and software to automate the provisioning and management of the lot. 

And here is where I fundamentally believe we can add real and demonstrable value. 

We are well versed with capabilities which far outweigh our competition in the components which make a virtualised datacentre possible.  Don’t believe me?

We have been involved with VMware since 2002 and won a Supplier Innovation Award in 2007 from BT

I’ve been developing collateral, running customer education events, and providing internal training around grid storage solutions such as IBM XiV, EMC Vmax, and NetApp ONTAP 8 internally for well over a year. 

We have more BMC, BladeLogic, and Tideway datacenter automation experience than anyone in the UK. 

We’ve automated the provisioning of data storage

We have skills in Cisco networking, virtualised I/O solutions such as Xsigo, and unified network architectures such as FCoE.

Put simply, we have demonstrable skills and expertise to advise and assist customers every step of the way from their existing traditional datacentre …and the low utilisation and high costs it undoubtedly entails …to a virtualised and automated datacentre with the high efficiency, increased utilisation, and lowered and predictable costs it promises.  What’s more, we have an equation in [ROI] + [CBA] + [DPB] = CSS which we use to quantifiably measure what our customers will reap by moving to the next step …and, in certain circumstances, we will agree with a customer to underwrite and gainshare the difference between their existing costs and the lowered costs our proposed solution has identified.

The end of history?  No, I think this is just the beginning …and Computacenter are ideally placed to help customers write the next exciting chapters.

As always, please don’t hesitate to contact me if you would like help in taking this journey.

Have a great weekend,

-Matthew

Click here to contact me.

And then there was one …erm, well OneConnect actually.

06/11/2009

I had the good fortune this week to have been invited by friends to fly a Boeing 737-300.  I arrived at LHR, went through my preflight checks, boarded the BMI 737-300, and proceeded to prepare for takeoff from runway 27R with flaps 5.  The copilot called out 80 knots …vee one called at 123 knots …I rotated the control column at 125 knots and we called vee two at 135 knots, cleaned up the gear and flaps and climbed to 6000 feet.  A quick right hand circuit around LHR and London, waved to Hatfield, and prepared to land safely at LHR.  I slowed down the airspeed to 136 knots and deployed flaps 15 …six miles to go and I slowed to 127 knots and deployed flaps 30 …four miles to go and we deployed flaps 40 …2500 feet called, down gear …see two reds and two whites on the PAPIS …80 feet to go …flare the nose and …touchdown!  A little bumpy, but a safe landing nonetheless.

Okay, so I didn’t fly a ‘real’ Boeing 737-300 …it was a full motion flight simulator operated by BMI at London Heathrow …but as simulators go this was about as realistic as you can get short of hearing the passengers behind you scream when you bank at 45 degrees over Canary Wharf.  From the time I walked in until the time they had to practically peel me from the captains chair I thought I had died and gone to geek heaven.  Those of you that know me know that I’m not just any run of the mill geek …nope, I’m an aeroplane technoweenie.  Perhaps not to the extent that you would mistake me for Colin Hunt, but Mrs. PL has sometimes come down at 02:00 to tell me that I’m not really flying to Japan its Microsoft Flight Simulator dammit now come to bed!  I know some men who feel they have to hide some magazines of a delicate nature from their wives …in our house, I hide my landing charts and cockpit video DVDs.

What does this have to do with data storage and protection?

I have known for quite some time just how difficult it is to fly a commercial airliner, although the experience of flying a Boeing 737-300 full motion simulator brought into even sharper focus the number of processes which must be mastered and continually monitored to ensure a safe flight.  Indeed, with the advent of low cost airlines many of us fly so frequently that we can easily forget to remember just how remarkable a thing it is that we can get on board a metal tube, have a seat, and fly through the air at several hundreds of miles per hour as I was reminded in this humorous clip.

However, it is rare these days that a pilot will fly a plane entirely ‘by hand’ as systems have been designed which can take off a plane, fly to a destination, and land again all on autopilot.  Why?  Well, the reasons are somewhat complicated …and most commercial pilots actually use a mixture of automation and flying ‘by hand’ …but, put simply, computers are far better at automating the speed and direction of an aircraft such that the human pilots can concentrate on more important matters such as not crashing, monitoring airspace to avoid collisions, monitoring the weather to move the plane to smooth altitudes and so on.

One of the more interesting storage products to have been recently announced is the Emulex OneConnect Universal CNAs.  Try as I might, I won’t ever be able to make host bus adapters [HBAs] nor network interface cards [NICs] ‘sexy’ …sorry Emulex! …but what the OneConnect CNA product is capable of is fascinating and deserves attention.

Firstly, the converged nature of the Emulex product is certainly interesting as it means that we can use the same card for both network and storage traffic.  This in and of itself would equate to demonstrable cost savings for our customers …you can essentially use one product where you would have previously used two …however Emulex doesn’t stop there.

The OneConnect products allow me to not only operate and leverage multiple network / storage mediums and protocols [Fibre Channel Over Ethernet a.k.a. FCOE, iSCSI, fibre channel], I can also offload these from the server to the chipset on the Emulex card.

If your eyes are glazing over just like they were when I was geeking out with the flight references above, here’s what that means in English …and why we should care.

The offload of protocols to the humble converged network adapter allows the server to do what it does best without having to carry the overhead of worrying about the network and storage protocols.  Just like we want pilots to worry about the important stuff and we’ll leave the highly repeatable tasks to computers and automation, this offload will allow the server to worry about the more important elements of the technology infrastructure.

How much more?  Well, Emulex claim that by using a OneConnect Universal CNA we can operate 20% more virtual machines on a server than we would otherwise.  Imagine our being able to ‘shrink’ a customer datacentre by upwards of 20% …and all of the management operating costs, power, cooling, etc. that go with it …by simply using another card which costs the same as competitive products and you can see why I’m excited about this product release.  Talk about Sharpen Your Business!

I know that we are all incredibly busy during this time of year and can often overlook which card we select when recommending a server and/or storage solution, but I would urge you to bear Emulex in mind.

As always, please don’t hesitate to contact me if you require any assistance in taking this journey.

Have a great weekend,

-Matthew

Click here to contact me.