Mikael Abrahamsson wrote:
On Mon, 1 Dec 2008, Deepak Jain wrote:
3) No one cares if the server farm is blast proof (it isn't), if the
connectivity in/out of it gets blasted (submessage: silos were meant
to deliver one thing, datacenters aren't in the same operational
model once they need connectivity to the outside world)
It's much easier to restore fiber connectivity in a time of crisis
than it is to source hardware manufacturered at the other end of the
world and have this set up properly. I do think there is value in
keeping the hw safer than the connectivity to the outside.
I bet the military or emergency services can establish a 10km fiber
stretch in a few hours. Replacing some telecom hw and set it up from
scratch would probably take weeks (I'm not talking about a single
router here).
Hi Mikael,
The speed with which fibre can be pulled will very much depend on the
available paths and other resources. It may be that the previous path
of the damaged fibre may now be blocked or otherwise unavailable such
that construction work is required.
As you say it is likely to be more difficult to recover from a problem
at a datacentre due to the greater potential for damage and diversity of
resources required. The point has already been made that not all
customers may be able to avail of site resilience due to the associated
cost, and so may be reliant on the one datacentre. In addition one
thing which I do not think has been mentioned is that damage to a
building may make the site unsafe and possibly injure staff; perhaps
causing planning, coordination and implementation of site recovery to be
considerably more complicated than simply replacing equipment.
Most customers would not be willing to pay extra to get hardened
datacentres, so despite the complexities of recovery Deepak is largely
right when he said that no one cares about blast proof server farms, at
least in the peaceful parts of the world.
Paul.
