HP duplicates deduplication
Redundancy cruncher for the big boys
SaaS data loss: The problem you didn’t know you had
Adds different enterprise deduplication product alongside SME D2D ones
Like EMC, HP is offering different deduplication products to different backup markets, with a post-processing Virtual Library System (VLS), based on Sepaton's DeltaStor technology, and inline processing D2D systems for small and medium enterprises.
Deduplication strips out redundant data from files at the sub-file level. HP is claiming it can provide a reduction in backup file size of up to 50:1; pretty ambitious. It assumes a standard business data mix and actual results will vary with data type, backup methodology and the length of time data is retained.
Inline deduplication is applied directly to incoming data. Post-processing deduplication is applied after incoming data from a backup run has finished arriving.
Dave Russell, a Gartner VP, said: “Deduplication technology is poised to transform the backup and recovery marketplace.”
HP's StorageWorks D2D 2500 and 4500 Backup Systems emulate up to 16 LTO tape autoloaders or libraries and can consolidate backup of up to 16 servers onto a single network-connected, disk-to-disk (D2D) device. These are not virtual tape library systems. They deduplicate data, using in-house HP Labs deduplication technology, inline, as backup data is ingested. HP calls this Dynamic Deduplication and notes that disk space is saved as no capacity needs to be set aside to hold incoming raw backup data.
Existing backup applications work with the products which are for small/medium businesses and branch sites.
Bob Wilson, VP and GM for HP's Storage Platforms Division, said: “HP continues to set new industry standards by investing in technology that delivers an ideal balance of price and performance for each customer segment.” The 'new industry standards' point comes from HP's belief that its new deduping systems are 45% more affordable than similar systems from Data Domain and the cost/TB of effective backup capacities.
The VLS product with its 'accelerated deduplication' is for enterprises, which typically have large backup data amounts, and ensures there is no compromise of backup speeds. HP claims it has market-leading scalability and performance by easily adding disks and nodes as needed. The company says that, compared to competing EMC/Quantum, Data Domain and IBM/Diligent products, the VLS leads in performance and capacity with 4,800 MB/sec - with eight nodes - and 1.3 petabytes of storage capacity.
These capabilities, coupled with reliable data recovery features, allow customers to create a data center environment that delivers 24/7 availability.
HP states that the data deduplication technology for the VLS and D2D enables customers to automate and remotely manage the systems with low-bandwidth replication. This provides data center managers the ability to back up data remotely without manual intervention, thereby reducing staffing costs. Utilizing disk-based backup for remote locations reduces overhead by consolidating tape hardware into a single site.
The HP StorageWorks D2D 2500 and 4000 systems are available now starting at a U.S. list price of $6,499. Accelerated deduplication is available for license with the HP StorageWorks VLS6600 and VLS9000 series and will be available in July. Licensing for the VLS6200 and VLS12000 models will be available in September. Pricing is based on physical capacity.
Copyright © 2008, Blocks & Files.com
COMMENTS
ManFromMars Alert!
No, wait... it's not in the comments section, it's the start of the article. Ow!
ZIP vs tar.gz
I have file collections that are 10 times smaller as tar.gz than as .zip
The difference is that zip compresses each file then starts fresh on the next one while tar creates a huge file with all of the contents of the sub files and then condenses the entropy in that huge file, which lets it take advantage of repeats between files and even in the file headers under tar.
Isn't this just called "stream compression" - like zip/lzh/etc?
LZH and friends are a family of "data deduplication algorithms"; huffman coding and all that jazz are based on well established principles of information theory, which is based around "de-duplicating" repeated data.
So, HP have a tape compression system. What's new about that exactly, apart from the name?
IT infrastructure monitoring strategies
Agentless Backup is Not a Myth
Top 10 SIEM implementer’s checklist
Steps to Take Before Choosing a Business Continuity Partner
Enabling efficient data center monitoring
