On Wednesday the 11th – at Dell Storage Forum in London – we launched the DR4000 disk-to-disk backup appliance, designed for SMB to mid-sized enterprise customers with small to moderate sized backup workflows. It’s always great to see a large company like Dell get a new technology into the market, where we can bring additional value to our customers. The DR4000 is especially sweet since it represents the second “Ocarina-enabled” product from Dell, and is a proof point that we are serious about investing in solving our customers’ backup problems.
Backup is a really interesting workflow, one that customers consistently highlight as an area with recurring headaches. Some of the issues we’ve heard:
- The cost and complexity of backup software and its licensing,
- The administrative hassles of managing tape,
- Major difficulty identifying performance bottlenecks,
- Managing the day-to-day requests for restore operations,
- Business problems like making sure the management’s expectations for retention and recoverability are in line with the protection solution they’ve budgeted for.
One of the most time consuming activities for the administrator is restore operations, especially if it involves transporting tapes back from a vaulting service. The response time to a user who “accidentally blew away his file” or a request to restore an x-employee’s inbox, is substantially improved if that backup is staged on disk. That’s not to say disk is faster than tape; it certainly can be, but you can also get a heck of a lot of throughput from a stack of LTO-5 tape drives! It’s when the RTO includes the time to drive a truck, find a cartridge, spool, shoe-shine and stream…that’s when the responsiveness to the end-user falls apart.
So we want to store as many backup images on disk as possible, but the capex and opex of a JBOD are hardly in line with the costs for tape. But it turns out that dedupe algorithms are a match made in heaven for backup workflows. If we can shrink your data 15-20x and thus capex/opex, then the staging of backup images can go from say 1 week to say 15 weeks, and the likelihood you’ll get a restore request for something over 15 weeks old is pretty low. Yeah! No restores from tape!
We built the DR4000 just for the needs of the SMB to mid-market business, where backup sets are 1-5TB, and customers want simple streamlined value that plugs-and-plays in existing backup environments. Generally this is a set-it-and-forget-it purchase for customers, with most of the administration of backups and restores happening through the backup software.
Here are some of the specs:
- 3 models; 2.7, 5.4, and 9TB usable capacity (before dedupe)
- Special in-band dedupe and compression algorithms coming from the Ocarina team
- Optimized replication: Only deduped data goes over the wire
- 2u chassis, RAID6 protection, 1G and 10G connectivity, GUI and CLI administration, NFS/CIFS/OST
- Initial certification with Commvault and Symantec (NBU and BE), with others coming soon
- All inclusive licensing so you get all the future enhancements for free
The DR4000 is one of what will be many investments in how we improve the protection solutions available to customers. In the future we’ll be adding more to this family, but where it gets interesting is when we apply Fluid Data principles to the roadmap, including ideas like more seamless integration with EqualLogic and Compellent, as well as Cloud services. More to come this year!
