It feels like quite a long time since I blogged just prior to my going on holiday with Mrs. PL and PL Junior, so please do forgive this Weekly View being a) somewhat late and, b) out of sequence. I am back and into the full swing of things, with the Weekly View commencing each Friday again from this week.
Many interesting things have happened since I went on holiday; here are just some that caught my attention … PL Junior started reception which left me wondering where the time goes, Lloyds made the front page of the weekend FT by piloting iPads, Paul Maritz [CEO, VMware] stated that ‘in 2009 organisations deployed more virtual machines [VMs] than physical machines for the first time’, and let’s not forget the $2.4b tussle between HP and Dell over 3PAR … with HP winning the tug of war.
Now, I have to make a confession here … I hate packing and unpacking possibly more than anything I can think of and will do just about anything to not have to do it … to wit, when Mrs. PL and I got married I suggested we throw our honeymoon clothes into suitcases dirty and then have housekeeping launder them when we got there. Made sense to me … no hassle packing, unpacking, and we’d have clean and neatly pressed clothes delivered to our room! Mrs. PL was less enamoured with this idea and offered some suggestions of her own for my packing which I’m still not sure are anatomically possible.
What’s this got to do with Data Storage & Protection?
I don’t like packing for two primary reasons; 1) I’m rubbish at packing and can never seem to get things packed properly, and 2) what looks like a large suitcase never seems to hold what it should.
Enter Mrs. PL who has a PhD in packing, bringing order to chaos and filling every cubic centimetre of our suitcases. Which got me thinking … packing is a three dimensional problem and what Mrs. PL does so exceptionally well is to bring three dimensional solutions to this three dimensional problem.
I believe the exact same thing … 3D solutions for 3D issues … is happening with technology generally and with data storage specifically.
1. Customers tend to reap only 40% utilisation from their storage infrastructures.
Customers want to get 100% utilisation from their storage infrastructures every bit as much as Mrs. PL wants to ensure she has used every cubic centimetre of a suitcase wisely. When it comes to storage inefficiencies, there are numerous reasons; fat provisioning, orphaned volumes, duplicate data, dormant data which hasn’t been accessed in days/weeks/months/years … even inefficiencies within the storage arrays themselves. The past two years have seen quite a consolidation of technologies such that the vast majority of tier one storage vendors have worked to introduce thin provisioning, data deduplication/data compression, automated storage provisioning/tiering into their storage arrays and offerings to increase utilisation. In essence, introducing what were once products in their own right as features into mainstream data storage. Why? Put simply, increasing utilisation decreases costs … and as customers continue to store more and more data they require the highest utilisation possible to avoid excessive storage costs.
2. Three dimensional problems require three dimensional solutions.
Virtualising a server, introducing dedupe into backups, thin provisioning a few volumes … each on their own are two dimensional solutions. Whilst two dimensional solutions will reduce costs somewhat, it is only when these solutions are coupled holistically into three dimensional solutions that true cost reductions both in the immediate and for future growth can be achieved. The Computacenter Virtualised Datacentre [VDC] solution is a three dimensional solution, seeking to holistically optimise the network, platform, hypervisor, storage, and automation such that OpEx and CapEx costs can be reduced by as much as 30% to 50% or more both in the immediate and for future workload creation and data retention.
3. 3D solutions, whether at the datacentre level or storage level, are business solutions and not technical solutions.
It is true that VDC is made up of technical components such as hypervisors, universal compute, virtualised 10GB Ethernet, grid storage, and automation … however VDC isn’t a technical solution. It is a business solution which seeks to reduce costs by optimising and reducing wastage between components, automating highly repeatable tasks such as server and storage provisioning, all with a view to aligning technology [IT] to business value. Why is this important? Put succinctly, when calculating Total Cost of Ownership [TCO], only 30% of TCO is represented in acquisition cost … the remaining 70% is OpEx. Many tools and technologies have focussed heavily on immediate return [e.g. TCO calculators, data dedupe] however the real long term cost savings remains in OpEx reduction. Helping an IT department optimise OpEx should return significant long term value and cost reductions, and we’ve spent a lot of time putting science behind this such that we can underwrite the business benefits as opposed to peddling the marketing hype. I’ve written at length about how and why VDCs help internal IT departments transition/evolve into internal service providers, so I won’t rehash that now but click here and here if you’d like to revisit these posts.
4. Using commoditised hardware in data storage and treating data storage as a commodity are not the same thing.
Grid or ‘scale out’ systems such as IBM XiV and EMC VMAX form the basis of the Virtualised Datacentre … and cloud computing. The secret to scale out systems is that they use commoditised hardware … Intel chips, SATA, SAS, and SSD drives … and use software to manage data placement and automated tiering. However, this isn’t the same thing as treating data storage as a commodity and buying ‘good enough’ storage at the lowest price to store yet more data as this strategy is, generally speaking, what leas to low storage utilisation and high IT costs. These next generation arrays represent the introduction of true business intelligence into the storage fabric and seeking to store data created as efficiently as possible from creation throughout the lifecycle. Indeed, without scale out storage VDI, service catalogues, automated provisioning, cloud computing, et al wouldn’t be possible at costs low enough to help organisations overcome the inertia of how storage has traditionally been purchased and allocated.
5. If three dimensional solutions are the future of technology and the key to significant cost reductions, why not introduce them into data storage directly?
A very good question, and certainly one that HDS have put to the market and customers with their recent Virtual Storage Platform [VSP] announcement. Whereas the arrays mentioned in item 4 are more intelligent than traditional storage arrays, it could be argued that one must connect them to other components to achieve a 3D solution. HDS would argue that their new VSP introduces 3D storage … scale out, scale up, scale deep … into the market for the first time.
I’ll be tackling the HDS VSP announcement in my next blog post , giving you my thoughts about how I think it stacks up to competitive technologies and solutions.