Network Bandwidth:

Many companies have 1 or at most 3 OC3's

• Our Installed Back-bone:
  • >9 (nine) OC-3's (@155 Mb/sec each).
  • >6 (six) OC-12 (@ 622 Mb/sec each).
  • >9 (nine) OC-48 (@ 2.4 GB/sec each).
• Transits:  4 (four) connections with GoodNet, AlterNet, PrioriNet, Nasa Ames.

Network Equipment:
Major networking equipment includes 5 (five) Cisco 7513 routers, four Cisco 7000 series routers, a Cisco Catalyst 5000 switch, two Cisco Catalyst 1900 switches and a Cisco CPW 1100 switch.  2 Cisco Routers on Standby.

Redundancy: Connections with Pacific Bell, MFS WorldCom, Brooks Fiber, MCI, SprintLink, NASA, and TCG.

Peering:  457 total peering agreements with companies like MCI, SprintLink, MaeEast, MaeWest, Nasa Ames, UUNET, InternConnect, GoodNet, PrioriNet, Chicago Ameritech.

• Grade-A Telco/Computing facility and 400 sq. ft power plant.
• 3-level fire suppression system
• 24x7 security
• 13 hour DC battery back-up system with Diesel Powered generator for +13 hr. extended power outages.
• 24x7 "manned" hardware monitoring and 24x7 software monitoring of all web servers using 5 minute redundant ping and probe tests from two separate alternative locations.

Webserver (s)

• Independent fully-loaded Pentium Web Servers running Microsoft NT/2000/2003 and IIS as well as RedHat Linux.
• Redundant Array of Independent Disks (RAID) storage on system boot cartridges as well as customer content.
• Multi-Point tape backup systems.
• Second catastrophic Failover site to alternative location in East Bay California.  Provides multiple physical redundancy. Standby "hot swap" servers at alternative location.
• Web servers are on the backbone/zero hop operation.

Emergency Recovery Plan

• Recovery Goals:
  • Mean-time to respond - <10 minutes
  • Mean-time to recover from a catastrophic failure - <4 hours
  • Mean-time between failures - >90 days
  • Uptime Percentile - >99%
• In the event of Network failures, Network traffic is automatically re-routed to several redundant connections.
• In the event of Power failures, DC power will continue for 13 hours notice until generator start.
• In the event of Hardware failures
  • Hard disk crash-Hot Swap RAID disk (<1 min).
  • Power supply/server failure-Traffic is re-routed to alternative backup T-1 location and Hot Swap Server is made active.
  • Router failure-Standby routers are installed.