How is geographically distributed database implemented?

As stated on subject line, how is this done for national corporation that has database that spread across the continent while only having a 'single' image of database that need to be realtime accessed by user that are so wide spreaded? Basically I'm referring to having multiple servers that is capable of reading/writing to the same database on distributed locations (e.g. couriers, airliners)? How is this being implemented today? To me, this is like one data image, hundreds of cities, thousands of people accessing it. How is it done???

Something like Oracle Parallel Server will do the job on the front-end, which can distribute the workload, but what about the back-end storage? With any storage technology like NAS/SAN, this is create a huge bottleneck because everybody is accessing the same database. Is it a problem today?

Ernest
ernestsiu_at_yahoo.com Received on Tue Mar 04 2003 - 15:46:00 CST