We have never experienced data loss due to a hardware failure.
Tandem Nonstop Systems are specifically designed to run business critical, continuously available applications.
They currently provide processing for:
- 95% of Securites and Commodities exchanges, inluding New York, London, and Hong Kong
- 80% of ATM cash transactions
- 50% of U.S. 911 calls
System Architecture
Our network consists of twenty-eight Tandem ServerNet S74000 CPUs. The CPUs are part of the latest processor family available from Tandem, now part of HP. The current network consists of 2 linked systems, which are on-line production systems. One system consisting of 16 CPUs is located at the main facility in Deerfield, IL. The other system consisting of 12 CPUs at a separate, remote site. In each Tandem system there is dual ServerNet fabric between processors. Both systems (facilities) are connected via ATM by two diverse T3 lines. One T3 line is routed east and the other west, forming a diverse fault-tolerant fiber ring. System processes are written in a fault-tolerant environment so that a CPU failure or other component failure should not interrupt system processes. In the event of a failure, each independent system has enough capacity to maintain our day-to-day business requirements. We have never experienced a data loss because of a hardware failure.
The Deerfield computer room is 7800 square feet and can easily accommodate many ServerNet computer systems and supporting hardware. The secondary computer room is 2400 square feet and can also accommodate many ServerNet computer systems and supporting hardware. In addition, each site has full power back up including UPS, diesel or natural gas generator(s), with multiple fiber access for both phone and Internet connectivity.
We maintain full T1 connections to multiple Internet Service Providers (ISP). We control our own DNS entries, which, by default, round robin across all ISPs. In the event of a failure of one ISP, we can modify our DNS entries to exclude that ISP in the round robin thus giving us a fault-tolerant “first mile” to the Internet.
Every attempt has been made to provide for the integrity of the computer system. Through the fault-tolerant capability of our system hardware and the provisions provide our customers with the greatest system availability possible. Each of our customer's can attest to our success in providing them an extremely reliable and dependable service.
RequestInformation