Migrating to the Spectre Engine

The Diver Platform 7.x release introduces a new data engine—the Spectre engine. After purchasing a license, you can upgrade and continue to manage your 6.x content using Workbench, as well as migrate your data from classic Models to cBases, and develop new 7.x projects and content to take full advantage of the power and efficiency of the new technology.

There are two overarching strategies that may be employed when adopting the Diver Platform and Spectre engine. Each presents its own advantages and challenges.

  • Migration In Place

    This option involves first upgrading the existing environment from 6.x to 7.x, leaving all existing processes in place, and then gradually developing Spectre-powered content in parallel with existing content.

    • Advantages: Something you can do right now.

    • Challenges: Severe strain on existing hardware. Strong temptation to not optimize and leave potential performance improvements on the table.

  • Parallel New Deployment

    In this scenario the existing servers are left untouched. New hardware is provided to the Platform specification, and content is developed from existing requirements as if for a new deployment. Considerable infrastructure can still be migrated and re-used, but the new technologies are leveraged out of the box to ensure peak performance.

    • Advantages: No disruption to legacy servers. Opportunity to optimize around new technology for peak performance.

    • Challenges: New hardware must be in place before migration work can commence.

NOTE:

  • For both strategies, new hardware is almost certainly required. The requirements of the Spectre engine differ significantly from those of the Model engine, and running both on hardware previously optimized for Model driven content only, will present as noticeable performance degradation. See Hardware and OS Requirements and then contact DI Technical Support for more detailed hardware requirements to ensure best performance for your environment.
  • DI has found that most customers will benefit from pursuing the second strategy: Parallel New Deployment.

The complexity of your current systems and projects using DI software will dictate how migration can best be tackled. The suggested method is not to convert everything at once, but rather to focus on one area that you determine could benefit from Spectre. It might be an application with lots of multitabs, big models, or where performance is slower than desired. Create a new 7.x project for this work and start your conversion. Note that clients such as DivePort can make use of both 6.x and 7.x content at the same time.

See also: