A company with a first-generation deduplication appliance is migrating to an HP StoreOnce 6200
Backup System. After migration, what will be the size of the deduplication block?

A.
4 kb

B.
8 kb

C.
16 kb

D.
32 kb

Explanation:

Deduplication Optimization
HP created a highly-optimized deduplication approach that introduces time- and space-saving
techniques. With the goal of eliminating the maximum amount of redundancy in its data inspection,
while also maintaining a small index to deliver the fastest performance, the company focused on two
components of its deduplication approach: an average variable chunk size of 4K, and a sparse index.
HP’s Adaptive Micro-Chunking uses variable-length data segments or chunks. The backup stream is
broken down into approximately 4K variable-length segments that are examined for redundancy
versus previously stored information. Smaller segments means there are more chunks and index
comparisons, which also means a higher potential to locate and eliminate redundancy (and produce
higher reduction ratios). Comparative solutions use block sizes that range from 8K to 32K. The
tradeoff with small chunk sizes is a greater number of index look-ups—which could mean slower
deduplication performance. However, HP Labs developed HP Predictive Acceleration technology to
maintain performance and reduce RAM requirements. By using a subset of key values stored in
memory, StoreOnce determines a small number of sequences already stored on disk that are similar
to any given input sequence—what HP refers to as sparse indexing. Then each input sequence is only
deduplicated against those few sequences. This minimizes disk IO and uses less disk and little
memory, creating more efficiency and enabling faster ingest and, importantly, restoration of data.

HP’s approach accelerates reads/writes, and delivers rapid ingest rates of up to 28 TB/hour.
Predictive Acceleration has enabled HP to require up to 50% less RAM than comparable solutions.