Why did Easter Island Implode?
So what do you do when your client says IT wants to cut
infrastructure costs by consolidating all of their separate repositories into a
monolithic one? On paper, using EMC’s sizing guide and empirical data (number
of objects and size of content) the consolidation may make sense. The following
is a partial list of areas that describe the dimensions of this decision:
- Performance and Scale Issues
- Political Fallout
- “Shared” Services
Performance and Scale Issues
The average size and number of content files is the
foundation of this analysis, plus amount of processing and messaging between
users. The EMC Sizing guide helps with the initial infrastructure logical
architecture. The guide’s recommendations can then be applied to future growth
of content and user projections. Content indexing and search, and security are
essential factors to the future of the user’s experience. If security, for
example, is complex, then search results may have a tough time processing
through multiple layers of identity and access control. If content files are
larger than anticipated then the indexing may take a lot longer when reindexing
for upgrades. For backup/restore, failover, and disaster recovery, complexity
is increased for the “routine” patching and upgrades.
Political Fallout
For medium to large content repository implementations there
are considerable risks for internal turf wars, for example, IT’s top ten
projects set the priorities of the IT’s services, however, if your business
unit is not on the top ten, good luck getting the resources you need to
complete your projects. So if the consolidation happens and a pool of money is
paying for all of the ECM solution, then it becomes a political tug of war to
get your work done. Downtime windows for even small fixes become bogged down
with “red tape” from a group that has been burned before and is consumed by
risk avoidance. Security can be an issue to, for example, if some content for
HR is extremely sensitive which mandates that HTTPS is used instead of HTTP.
This slows down content transfer and overall performance which in turn requires
more hardware to speed up.
“Shared” Services
The concept of sharing is that everyone has an equal portion
of resources to get their work done. This model works for IT because they need
excuses to fall back on when projects are delayed. The portfolio model (like
ITIL) fits neatly into shared services in that work is identified and
prioritized, however the small projects get constantly pushed back and they add
up. The business becomes frustrated but has no leverage to fix the situation.
The success of IT should be measured, not by the execution of its big projects,
but the attention it gives to the many smaller projects that usually languish
and never see the light of day.
