We own our own Network Operations Center near
the
Big Chicken.
Our backup generators run on natural gas out
of the pipe. So, it won't run out of diesel or gas during an
extended power outage. We also maintain rack space at 56
Marietta Street in Atlanta.
Net2Atlanta is multi-homed
to the Internet via diverse paths with redundant links. In our
case, that means that one of our upstream providers are accessed via leased
copper wires and the phone company. But three others are accessed via
underground fiber. If a Bell Central Office has problems, the fiber
still works. If the fiber gets cut, the copper still works.
The fiber is also on a
SONET ring. That means that it takes two cuts
in different places to kill the fiber connection! Now that is fault
tolerant! And we have a second fiber network from AT&T to back up
the AFS SONET ring.
We are extremely well connected to the
Internet. Currently, our upstream providers are Sprint, Level3,
Global Crossing and Cogent. And, we are peered at
AtlantaIX to a
number of other ISPs big and small, including Earthlink, Akamai and the Georgia
University System.
We use BGP4. Border Gateway Protocol is what all the big
ISPs use to advertise their networks over multiple links and make them
fault tolerant. If a link
goes down, the packets find another path automatically. So, if a
link dies between us and our upstream providers or anywhere on the
Internet, your traffic still gets through! Some of our
customers are also multi-homed with BGP4.
Of course, we
monitor all of our main links as well as every one of our dedicated T1 and
DSL customers to
make sure they are always on. And we make the call if they go
down... often before the customer even knows they're down. We
also look for hacker activity.
We can set up our customers with redundant links
also by combining Fiber, Fixed Wireless, T1, DSL and ISDN
failover systems. For more fault tolerance, we can connect a
customer to both of our locations so that a disaster at either will not
take the customer down.
Read more about
this
We are also a computer
repair organization, so we stock
parts and spares to replace or repair any failures in our equipment or our
customer's rapidly. We have at least one plug replaceable spare for
every computer, router and switch on our network.
The result of all this effort is a highly
reliable, proactive system and satisfied customers.
Sure, all this sounds
obvious. Any responsible person would do the same. Then why
don't all of our competitors do this too?