Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

HP: Not All Quiet On The Western Storage Front: Page 2 of 2

Conceptually, deduplication is very simple. The system’s task is to remove duplicate data, such as from a current stream of backup data, as compared to not only the stream itself, but what already exists in the backup disk repository, where all previous backups are held. The goal is to keep one copy of each unique piece of data. The benefits are two-fold: 1) to reduce the amount of capacity that is needed to store data, and 2) to reduce the amount of data that has to be sent over a network when implementing an automated disaster recovery strategy. With disk-based backup systems gaining popularity as the first line of backup-restore defense, deduplication is considered more and more essential for any D2D package.

Although pioneers in any new technology have a first-mover advantage, later entrants into the space are by no means locked out. The reason is that later entrants are able to take a fresh look at the problem that needs to be solved and the limitations of the initial solutions, and are thus able to avoid, at least in theory, the mistakes of earlier entrants. HP Labs feels that it has taken advantage of the extra time to come up with three important innovations

  • Automated, intelligent block size optimization — HP uses a “variable chunk” size approach; chunks being blocks of data that have to be managed during the deduplication process; StoreOnce leverages what HP calls a TTTD (two threshold, two divisor) algorithm to reduce the average chunk size and variability between chunks; the advantage is greater deduplication efficiency .
  • Sparse indexing — An index helps to keep track of all the places where a chunk has been used; each time a new chunk is read, a check has to be made to see if the chunk already exists; the less that the deduplication software has to read in doing this, the more efficient the system performs; HP Labs has developed what it calls a sparse index that matches the new data to old data in a fraction of the time of other similar technologies.
  • Smart data and index layout —Traditionally, data becomes fragmented as it is deduplicated, and that leads to very slow read/recover times; HP's data/index matching scheme keeps more chunks local, which helps reduce fragmentation.

    All in all, HP StoreOnce seems to deliver a full, robust data deduplication software solution.

    Knowing other deduplication vendors, I am sure that their respective competitive analysis groups have examined HP’s claims for StoreOnce inside and out and have positioned their products accordingly. The comparative advantages and disadvantages of other products is not the point, however. The point is that HP appears to have a competitive product with StoreOnce and that means that it is in the hunt for D2D sales with a product which can state the company’s case for customers.

    It is a good sign for HP to be a little more proactive even while it waits for the “i’s to be dotted and the t’s to be crossed” on its broader strategic initiatives, which can then be translated into targeted competitive positioning. In addition, it is good to see HP Labs sparking what will likely be a commercially viable enterprise solution. Life has to go on and that means putting revenue bread on the table today as well as preparing tomorrow’s menu.

    At the time of publication, HP is not a current customer of David Hill and the Mesabi Group.