Untitled Document
The service-oriented architecture (SOA)/software-as-a-service (SaaS) movement
is slowly transitioning from hype to passe in the blogosphere -- a sign of growing
maturity in both intellectualization and practice.
Now, the spirit of service innovation moves to a new arena: data management.
Are you ready for "self-service cloud sandboxes?" This may sound like
child's play, but the reality is that it's very serious business. These sandboxes
may well be the new laboratories where knowledge workers vet ideas that transform
business.
While data management in the cloud conjures up common issues such as trust,
service level agreement negotiation and data asset management, the sandbox concept
is a bit of a throwback, but with a twist. Remember when everyone was buzzing
about data marts? However, the cost of integrating marts was a bone of contention.
That perception may be changing.
Toss in virtualization, and today's data marts employ views and aggregations
from an enterprise warehouse to create what can be considered as virtual, private
data clouds, or sandboxes, to build and test ideas expressed in the form of
analytical models.
Such virtual data marts can be created fast, leading to the call for self-service
capabilities so that business end-users can make their requests and begin analyzing
their data with minimal enterprise warehouse support personnel and minimal disruption
to the ongoing operations of the warehouse. End-users can even add external
data to their sandbox and use the mining tools they prefer, akin to the way
children bring their own pails and shovels to their sandbox adventures.
And, when end-users are finished "playing in the sandbox," they've
either constructed a "sand castle" in the form of a useful analytical
model that's ready for prime time deployment, or they just walk away and let
the next wave wash off whatever was there -- even if there were partial sand
castles that just never took shape. Then, the set of next sandbox players get
their beachhead allocations, and the process begins again.
For awhile, it's going to be difficult to keep all of the terminology straight.
Some refer to the sandbox as an "elastic data mart," for example.
In addition, the term "agile analytics" is popping up to reference
the fact that the process relies on perishable resources -- the temporarily
allocated space, cloud, appliance or extra (reserved) capacity that the current
enterprise warehouse isn't using up.
-1-