How do we make High Availability Datacenters?

Hi Guys,

How do we make High Availability Datacenters?. If you have any ideas let me know. I am just trying to find out.

Get many backbones

[whcdavid]

Here are the some of the main thing you should consider.

1) Power distribution system (main power and ups
power) maximum duration of ups power

2)Datacenter physical security (Location, building
materials and Basement or floor level)

3) Building access and its security system, fire and
bio-gas indicators and much more

4) Number of backbone providers and datacenter High
Availability routing and backbone feeds. Most of the
Major telecom providers (UUNET, MCI, SPRINT, TELUS,
ALLSTREAM, Level3, Quest, AT&T, BT,FT,SWISSCOM and much more) They
will have major backbone lines in the country. Most of
the time they are reliable, but not guranteed. If a
datacenter has 2+ backbone providers then much
reliable.

Here are my suggestions. Am I right guys?

Thanks
David.k

seems right to me!

[SiteSouth]

A well designed emergency action plan. Clear guidelines for your system admins to implement the action plan.

Redundant routers.

Datacenter High Availability (HA) Program consists of proactive and reactive support and services. The proactive services include: assessments, change management, and configuration audits. The reactive services include 24x7x365 support, partner notifications and triggers. There are many things make high availability datacenter.

David has it pretty much covered, If your looking to create a high availability datacentre, you have to think from the what-if style of logic...

What if my power fails?... Do you have a backup generator?
What if there are routing issues... Do you have multiple backbones?

Once you've got all those bases covered, Start adding extra little features. Stuff that the big companies offer (rackspace.com) etc...

Hope that helps

This sounds like a plan I am also working on. If your looking for a partner or something feel free to contact me. Thanks. I hope to hear from you.

Joe Thompson

You must remember that you need internal network redundancy as well. If you have a single router and that fails, it does you no good if you have multiple carriers. If you have a single core switch and it fails, it does you no good if you have multiple routers and carriers feeding into it. The goal is to have no single points of failure, anywhere. This includes not getting connections over the same fiber, which many providers do. They enter a facility such as Equinix, get access to many carriers, and then haul it out of the facility over a single fiber pair. What if something were to happen with that single fiber pair?

This is a good point, you need to plan for anything that may happen.