Number-crunching in the Cloud
Wave BY:BY to old-school data analytics
Big Money from 'Big Data'
A McKinsey study last month claims that this Cloud downpour, "Big Data" in the jargon, presents a huge opportunity for businesses. And as we know, every opportunity is a problem. But one that, as Suedkamp suggests, can be solved – or at least ameliorated – by the Cloud itself.
"Cloud computing is clearly enabling companies to reinvent the way they do business," he claims. "From an IT perspective the Cloud delivers services faster, because of automation and standardisation. It enables integration across Clouds as well as enterprises with pre-built templates. And increases efficiencies with virtualised consolidated scaleable resources". And Clouds can be used to crunch all this "big data".
Clouds come in different flavours, from the purely private Cloud run internally using the enterprise's software on its own hardware, through a variety of third-party-hosted but still private Clouds, all the way across to the public Cloud accessed through the general internet renting its services for commercial use by the hour on a credit card.
Sehgal judges the wholly public Cloud "...appropriate perhaps for some analytic apps. But the downside is security and difficulty of integration with the enterprise's existing services."
IBM currently favours a modified version of the public Cloud, "Shared Cloud Services", essentially a stack of standardised SaaS solutions that are rented to enterprises connected by VPNs over the Internet.
Suedkamp powerpoints this with some Forrester-derived slides where the opex versus capex argument tops the list of SaaS adoption drivers at 72 per cent, with lower overall TCO coming in at 68 per cent and at third place speed of implementation and deployment at 54 per cent.
On-prem may be cheaper
Then by way of balance there's more Forrester data about the possible downsides of SaaS. The percentages here are much smaller, though. Security concerns are number one at 48 per cent, and oddly TCO is in there too, coming third behind the challenges of integrating SaaS into existing applications (39 per cent). Apparently 34 per cent of those surveyed, in Suedkamp's words, "would argue that the total cost of ownership may be higher than an on-prem solution".
Who was being surveyed? These data may come from the Forrester report "SaaS Adoption 2010: Buyers See More Options But Must Balance TCO, Security, And Integration" in which "Forrester recently interviewed more than 1,000 enterprise software decision-makers to find out their investment strategy for 2010". But Suedkamp doesn't say, and it would cost you $500 to be sure of that.
Which may set you thinking. It did me. We're sitting here being talked through slides – how familiar is that? – with precise figures attached to somewhat less precise concepts, derived from reports, refined from summaries of raw data boiled down by machines, somewhere in the Cloud perhaps. It's analytics, being used to make us feel like informed decision-makers. And in this case to sell us analytics.
What do we really know about the provenance of this kind of data? And of the methodology used to crunch it? Dare we hope that what's going on behind the scenes – whether it's Forrester predicting trends, or our own business feeding information to our decision-makers – is far more rigorously audited than very similar processes on much the same machines that only very recently rolled up millions of sub-prime mortgages into slick collateralised debt obligations?
Or are we perhaps entering the age of the zettabyte spreadsheet? ®
Sponsored: Hyper-scale data management