So far we’ve covered 3 of 5 data deduplication best practices: considering the broader implications of deduplication, learning what data does not dedupe well, and why obsessing over space reduction ratios isn’t necessary. In this post, we’ll discuss two final deduplication best practices.
If you’re backing up to a virtual tape library (VTL), don’t use multiplexing. Even if your deduplication solution can de-multiplex data, consider turning this feature off. Often a carryover practice from writing to physical tapes, multiplexing data merely wastes computing cycles— cycles that could otherwise be used to dedupe your data faster. For example, instead of multiplexing ten backups to two virtual tape drives, create twenty virtual tape drives and turn off multiplexing.
Before selecting your deduplication solution, try to pilot several deduplication systems in your environment. While current vendors offer many good solutions and various deduplication approaches, you may also find some products with real limitations. Only by comparing multiple products can you best determine the optimum approach for deduping your data, whether it’s inline, post-process, target-side, client-side, via backup software, etc.
Common challenges of deploying a deduplication solution involve problems related to performance, increased complexity of management, and proliferation of deduplicated data silos. To avoid unnecessary complications, first ensure ease of integration into your existing environment and get customer references in your industry. Take time to understand the vendor’s roadmap, but test everything. Once you’ve selected your data deduplication solution, make sure you follow the best practices suggested by your deduplication solution vendor.
