In the February, 1997, issue of NETWORK VAR, Roger Addelson outlined a whole series of y2k problems that face networks. Some systems will not roll over. Others will have problems with time stamps. If the host operating system has a date problem, the whole system will.
* * * * * * *
According to the Gartner Group (Stamford, CT), 90 percent of applications will be affected by the date problem, and systems will crash if the century problem is not solved prior to January 1, 2000. Twenty percent of business applications will fail because of incorrect date applications. . . . Although this Gartner Group study focuses on legacy mainframe systems, similar estimates can be made for networked PCs as well (p. 27).
In the network arena, many operating systems for bridges, routers, and gateways are proprietary. Although standards are emerging among these operating systems, specific protocol implementation tends to be proprietary to each vendor. These protocols are extremely time-sensitive; only the most recent routing and bridging information packets are kept to update the routing tables. Older information is discarded. If the routers and bridges have the known time/date order anomaly in the year 2000 problem, and it is not identified and rectified, enterprise network traffic can grind to a chaotic halt at midnight, 12/31/99 (p. 29).
As you may have noticed, I raised more questions than I provided answers. That is the problem -- there are more questions than answers. Systems are so diverse and interconnected that each system as a whole, as well as each component of each system, must be examined. Source code must be examined line-by-line. System interdependency and interaction must be tested transaction by transaction and field by field. . . . A single solution is impossible because of the interdependency between hardware platform, operating system, network, and application. The number of combinations and permutations are [sic] astronomical. . . .
The time to start was last year (p. 30).