|
|
|
|
|
|
|
|
|
|
|
|
Both in response to this and in response to concerns from within the corporation, UCAR has established a "Y2K Task Group" to develop a plan for addressing problems that may arise due to the transition into the next century, and SCD has played a major role. The Y2K Task Group is headed by Steve Dickson (UCAR), and its membership represents UCAR as a whole. The Scientific Computing Division is represented by Tom Engel.
A preliminary UCAR Year 2000 Plan was drafted during the first quarter of FY1998 and submitted to the NSF for review. The UCAR plan follows the outline of phases and the key processes within each phase as promulgated by the U.S. General Accounting Office in their "Year 2000 Computing Crisis: An Assessment Guide" (GAO/AIMD-10.1.14). The UCAR plan has five phases corresponding to those identified by the GAO; the phases and UCAR's general timeframe for each phase are:
This plan attempts to ensure that all UCAR-owned and/or managed systems are "Year 2000 compliant" well before 1 January 2000. Compliance is defined (by the GAO) as "the ability of information systems to accurately process date data from, into, and between the twentieth and twenty-first centuries, including leap year calculations."
During FY1998, SCD continued assessment and began renovation and validation of certain key (a.k.a. "mission critical") systems, in particular the supercomputers and NCAR Mass Storage System (MSS). All supercomputers were upgraded to Year-2000-compliant operating systems and software product sets during FY1998 (with the exception of those systems to be decommissioned before 2000). The Mass Storage System was made ready for planned Y2K testing, which is scheduled for early in FY1999. All SCD divisional servers and systems were assessed during FY1998, numerous systems were decommissioned, and remaining systems will be upgraded to Y2K-compliant system software early in FY1999.
SCD has published its Year 2000 plans as a set of working documents. This set of documents is updated frequently as assessment, renovation, validation, and implementation activities proceed. Interested parties are advised to periodically check the SCD website for updates. In addition, SCD has published an inventory of systems it is responsible for administering and maintaining, along with their "state of readiness." This web page is also updated regularly as status changes.
The High Performance Systems (HPS) section of SCD is currently engaged in executing its plan to address the Y2K "problem" within SCD and NCAR. HPS is working with other sections of SCD, particularly with the Network Engineering and Technology Section (NETS), to develop and execute plans and procedures to ensure that no mission-critical systems will fail to perform properly into the next century.
In brief, SCD's plans for addressing the "Year 2000 Problem" are to initially work with computer system vendors to upgrade all existing systems to be "Year 2000 compliant." In most cases, this requires an upgrade of the operating system and the vendor's software product set. Once this is accomplished, SCD will perform isolated single-system testing and evaluation, followed by multiple-system interoperability testing. In addition, SCD-developed systems, such as the NCAR Mass Storage System, NCAR DataVision (NCAR Graphics), the MASnet/Internet Gateway Server (MIGS), Internet Remote Job Entry (IRJE), etc., will be tested to assess any Year 2000 vulnerabilities. Corrective action is planned for any deficiencies identified. Specific plans are being drafted within HPS for each major service area: supercomputers, mass storage systems, distributed services, and desktop systems. The Network Engineering and Technology Section (NETS) is also drafting a plan to ensure the soundness of the NCAR/UCAR network infrastructure through the transition into the next century.
For systems that are not explicitly tested or evaluated and determined to be Y2K compliant, special plans will be made for the January 1-2, 2000 weekend and subsequent week(s) to assess and monitor the behavior of those systems and resident applications. These plans will include keeping the appropriate staff on call and available to debug and repair failures in mission-critical systems or subsystems if they occur.
In addition to the assessment, evaluation, remediation, and testing prior to year 2000, SCD plans to actively monitor all production systems over the crucial January 1-2, 2000 weekend. SCD's Operations and Information Support section, along with the High Performance Systems and Network Engineering and Technology sections, plan to augment our 7x24 monitoring and on-call procedures and staff over that weekend and through the succeeding week(s) to respond immediately to any unanticipated problems. The extent to which this monitoring is augmented, and decisions on whether production systems will be shut down prior to, and restarted after, the 2000 new year depends on the evaluation and testing conducted on these systems (both by SCD and the system vendors) in the upcoming months.
The design was to provide a small, integrated system resembling that which SCD runs operationally to provide an environment in which Year 2000 testing could be performed without impacting the production systems and networks. Unfortunately, funding was not available for the acquisition of equipment to deploy such a testbed. SCD is continuing with its plans for Y2K testing and remediation during FY1999, but it will be using existing systems (sometimes taking production systems down for extended periods of time) for that testing.
|
|
|
|
|
|