Migration and transition
One of the most difficult aspects is the transition process itself, especially with very high availability multi-site clustered systems with extensive storage and network infrastructures. The problem is more one of complexity than scale. Most migrations will be relatively straightforward. Some will be exceedingly difficult. A relatively small scale but highly complex and tightly integrated system with exceedingly demanding availability requirements will be one of the most difficult kinds of system to migrate successfully.
Here be dragons - "Hic sunt dracones"
From the earliest times, map makers used to depict scary monsters and other mythical creatures as inhabiting unknown territory. The Hunt-Lenox Globe in the New York Public Library carries the Latin inscription "Hic sunt dracones" (here be dragons). The phrase has become synonymous with entering diffioult and unknown territory. Guides who knew how to cope with such territory were highly respected and sought after, because they greatly improved the chances of making it back in one piece.
Most OpenVMS systems have the applications and operational processes tightly integrated with the operating system, hardware platform and surrounding infrastructure. These are time consuming and difficult to understand. They will be inherently difficult to migrate. XDelta will be your guide in such difficult territory.
The scale of the migration problem:
A number of factors contribute to the overall complexity and dificulty of any systems migration project:
- Most systems have evolved over many years: PDP11 to VAX to Alpha to Integrity is not unusual.
- A range of languages: Basic, Fortran, Cobol, Macro32, C, C++, Pascal, Ada, Java, DCL, etc.
- A range of layered products: Rdb, DECforms, ACMS, RTR, DECset, ABS/MDMS, etc.
- A variety of workloads: Interactive, Batch, Real-time, Backups, etc.
- Set up in different ways: Naming conventions, Programming styles, DCL usage, etc.
- People have moved on: Knowledge, Documentation, Design techniques, etc.
High performance systems:
These pose a particular challenge with severe constraints on throughput and responsiveness. High IO write rates to storage are particularly difficult to achieve.
High availability and multi-site systems:
These pose a particular challenge with severe constraints on permissible outage times. Multi-site operation brings additional complexity with long-distance data replication and the associated latency and distributed locking effects.
Those running large and complex systems with both high performance and high availability requirements have the greatest problem. In many cases the transition process itself is one of the major constraints on the migration design.