Part 4 of 5: What about DRR (Data Reduction Rate) in an as-a-Service model?

IaaS, PaaS, CaaS, XaaS (Everything as-a-Service) and of course Storage-as-a-Service…  You can’t get away from a dramatic shift in the IT market place to deliver services rather than just products.  In my business this year, 24% of my deals are as-a-Service.  To  put some more context on that, precisely 50% of my teams revenue comes from some of the largest banks in EMEA so draw your own conclusions.  Banks understand finances, and my banking customers are buying their storage “as-a-Service”.

Anyway, back on topic, so what does Storage-as-a-Service have to do with DRR?  Well, interestingly, when your vendor (or service provider) owns the storage assets – who takes the risk on not achieving high DRR, and who takes the reward when you get great DRR?  To help explain this, I’ll pose a few questions to my colleague Abe Barnes (LinkedIn).  Abe lives and breathes Storage-as-a-Service models, and quite likely knows more on this subject than anyone in the business!

“Abe,  If I buy Storage-aaS – do I need to worry about DRR?”

If you are truly subscribing to a service, then you shouldn’t need to. However, the fact is that not all Storage as-a-Service models are the same. If your storage as-a-Service provider is telling you that their model “gives you the benefit of DRR” then it means that they are committing you to a piece of technology – an array of a particular type or a specific amount of capacity.  That type of model does not have the level of flexibility, agility, or cloud-like experience that most customers have come to expect from true services.

“Thanks Abe, but what about capacity?  In an aaS model, how can I be assured that I will always have enough capacity to meet my businesses demand, and are you going to make me pay for capacity that I’m not actually using?”

A good storage service should be based on service outcomes and assurances, and not limited to technology specification or features (such as a specified amount of capacity). Having capacity always available should be table-stakes in a storage as-a-Service offering and should ideally flex in line with usage. In a truly modern storage as-a-Service model, your vendor or service provider should provide you with assurances on availability, capacity, and performance, and should have the ability to adapt the underlying technology accordingly and without you making any additional commercial commitments, you just pay for what you use (and watch out for those models that say they charge you for usage, but they really just adapt their rates based on how well the DRR is performing, because you guessed it, they’re committing you to technology and not to a service experience).

“Sounds too good to be true – there has to be a catch?  Is it outrageously expensive or something?”

Actually we find that when it’s done right, consuming storage as-a-Service is more cost effective than purchasing storage assets, managing and maintaining them.  It’s important to remember that the charges associated with storage services are not just the cost of the hardware spread out over time, but tend to include maintenance, upgrades, and a degree of service management designed to plug in to your own ITSM ecosystem – to make sure those service assurances I mentioned are delivered.  When it comes to planning and comparing costs of storage consumption models, think about the business value of being able to consume in a flexible way, only pay for what you use, and holding your vendor or provider to account for the capacity, performance, availability, and overall health of your storage estate.

A great roundup from Abe on why you have no need to be concerned about DRR when you’re buying your Storage in an as-a-Service model.  I liken this to the idea of just stating the effective capacity on a box and leaving the tech magic up to the vendor.  In Abe’s world – you go one step further, and you don’t even pay for all that storage that you’re not actually using.  How often do customers make the mistake of dividing the cost of an array by the capacity (usable or even effective) to arrive at a per TiB  cost?  I’ll tell you- all the time !!!  WRONG WRONG WRONG!!! 😱 – that only works if you write to every one of those bits of silicon!  Most customers are running arrays at 50-70% capacity – so those $ per TiB values could be as much as double what you thought they were.

Watch out for another blog series on Storage-as-a-Service where I’ll ask Abe to explain in more detail the real value of this model to your business.