by Bruce Kornfeld, Lead, Dell Storage Alliances — September 07, 2010
Looking back in time, backing up data from desktops and servers to tapes made sense. There are just too many things that can go wrong with an IT system; you’d be foolish not to back up your data. Having that insurance policy sitting on a shelf or at an offsite location was critical. If an “event” occurred, the clock starts ticking and it is IT’s job to bring their organization back online. How long? A few hours? A few days? A week? Looking toward the future, using tape as an insurance policy for bad things happening in the datacenter just won’t work for much longer. Data sets are too large, pipes from tape back to disk are too slow, and the time it takes to physically find tapes, ship tapes and restore from tape just doesn’t cut it anymore. The pace of information flow and business change will be too fast for restoring from tape in the future. (Are we already there today?)
Alternatives to tape backup
There are many technologies out there offering superior ways to ensure that operations can be back up and running shortly after any “event” or to avoid downtime completely—based on budget, of course. Many forward-thinking IT departments back up data to disk so that restores happen almost instantly. No more searching for the right tapes and waiting for tapes to spin in order to get back up and running. Some of the techniques that create backup sets on disk available today include:
- Increase the snapshot frequency on primary storage—manage those snapshots with the disk vendor’s user interface or a backup application. The data is already there ready to be used. These snapshots can also be sent to another physical disk system to provide extra protection.
- Use backup software (the way you always have) and write to disk instead of tape—a pretty common practice these days. The data could also be de-duplicated within the target disk. However, I suspect many are still ALSO writing the same data to tape—is this really necessary?
- De-dupe the data at the server and then send to disk—another method that is gaining momentum because less data needs to be sent to the backup disk—less to manage.
- Send backup sets to another location for disaster avoidance—bandwidth is a lot cheaper than it was 10 years ago and if you’re sending snapshots or de-duplicated backup sets the amount of disk needed in the remote location can be price competitive with sending, storing and managing tape after tape after tape...
What about accessing “old" data?
All of this discussion was about backup—being ready in case something bad happens. What about accessing all the old data that doesn’t really need to be accessed frequently? SEC filings from 3 years ago. Engineering drawings from the last product revision. Surveillance video. The CEOs keynote speech from the customer event 2 years ago. This is where an active archive comes in.
As a disk vendor, we’d LOVE to think that every organization would store their data on their primary storage disk forever. But that’s just not reality. Automatically tiering data within your primary storage is a great thing. But now there are techniques to create archives on secondary disk and tape systems so that your organization can have fast access to the information they need. It might appear that I’m suggesting that data never goes away. It just moves from fast disk to slower disk to tape over time—staying online and available the whole way. That’s the whole point. Archiving software is intelligent enough these days to help IT departments decide how long data needs to stay available so managing this process is getting easier. The key difference here is that these archives can now be on tape systems. Tape for archives—YES. Tape for backup—NO. Maybe the industry as a whole isn’t ready for this yet—but more and more will be moving there over time.
by Liem Nguyen, Director of Communications and Social Media, Dell Storage — September 02, 2010
All the buzz at VMworld 2010 has been around virtualization. I had a chance to catch up with Bruce Kornfeld, VP of business development and alliances from Compellent and David West, VP of marketing and business development from CommVault to talk about the partnership between the two companies, how the two technologies work together, data storage efficiencies, and data management and protection.
It wasn’t all work and no play for Compellent and CommVault at VMworld 2010. The two companies hosted customers and prospects at Tres Agaves restaurant on Tuesday, August 31 to unwind after a long day at the show and enjoy some great food and great company.
Watch the video to learn more about the way Compellent and CommVault technologies work together.
by Compellent Technologies, — May 06, 2010
8:17 am - Bruce Kornfeld and Marty Sanders make an appearance to start off the day talking about Compellent’s product road map. We welcome feedback and help prioritizing what you’d like to see next.
8:19 am - Fluid Data is new, but the underlying architecture is not. We’ve been using it all along, this is just a new way to describe Compellent’s feature set and capabilities.
8:20 am – We’re shipping RAID 6, and we also have the ability to take replays of multiple volumes at the same time (called ‘consistency groups’). Other Storage Center 5 features include Virtual Ports, Portable Volume, server mapping and standards-based space recovery.
8:22 am – Audience Response question: Please describe your replication use today:
- None – 33%
- Server-based replication – 18%
- Compellent storage replication – T3 or less – 24%
- Compellent replication – OC 3 – 12%
- Compellent replication – OC 12 – 5%
- Compellent replication – >OC 12 – 8%
8:24 am – Portable Volume reduces bandwidth requirements and saves time with automation and encryption. It’s an easy way to bring initial synchronization up.
8:25 am – Audience Response question: The fastest storage growth in your environment is coming from:
- Files/unstructured data – 51%
- Email – 12%
- Database & custom applications – 37%
8:27 am – zNAS: we’ve built a file system based on ZFS, runs on the same platform as Windows storage server. Automatically configures Storage Center on the backend.
8:29 am – Live Volume allows for planned migration – it’s non-disruptive. We’re exploring a lot of roadmap options going forward with this product.
8:30 am – We have a great platform based on Fluid Data – our goal is to move that platform to new places in storing data, to have it work seamlessly with new technologies. It’s all based on this same platform, no forklift upgrade necessary.
8:31 am – Compellent Metrocluster will offer easy, practical and affordable enterprise data protection; it’s the next step from Live Volume. Operations keep running, even if one side has an outage; the RAID and system redundancy are built into one system.
8:37 am – Standards based Space Reclamation – we’re one of the first storage companies to support SCSI SBC-3. As Operating Systems start picking up this technology you’ll have the capability to reclaim disk space after files are deleted by any OS.
8:38 am – We’re excited to come out with the Series 40 Controller. There are a lot of advantages – we’re adding another slot (7 total), all will be PCIe. The performance out of this box is just incredible, and we’re looking forward to better and better performance.
8:41 am – We’re very excited about this next generation scalable system – moving further into the Enterprise with the same platform. New capabilities will include: Series 40, larger cache (no more batteries), extreme scalability, more connectivity, and de-dupe options.
8:44 am – Dynamic Block Deduplication – we’re still working on this, but think about what we do from a Fluid Data standpoint: we put the right types of data on the right types of drives at the right time. Now, when we’re doing this, we can take a look at the blocks and compress if possible – it’s just part of the natural flow of data progression. Good show of hands when asked who’s excited about this technology!
8:46 am – How many people are using SAS technology from Compellent? Very few – probably because you’ve been using FC and SATA for some time. These are continuing, but SAS is the future of the back end of the Compellent Storage Center. It’s higher performance – we’re going to 6 Gb later this year. 2.5 inch drives will also be shipping.
8:50 am – The next phase of Storage Center is having the ability to take multiple storage centers and have volumes exist partially on each system. As replays age, have them automatically migrate to less-costly storage. Flexibility with the high-availability environments.
8:57 am – We’ll be doing more work on application integration, with partners like VMware (SRM, vSphere API with space reclamation), Microsoft (PowerShell, SCOM, DIT-SC (Dynamic Infrastructure Toolkit for System Center), Citrix (StorageLink with replication), Commvault (replay integration with Simpana), and Oracle (RMAN integration). Also Enterprise Certifications with IBM (SVC, VIOS, DB2, Tivoli), HP (Storage Essentials), and Symantec (NetBackup, Array Support Library – Linux, Solaris, Windows, AIX,HP-UX).
9:01 am – Audience response: Which of these Alliance Partner projects are you most interested in?
- CommVault snapshot integration – 27%
- Oracle RMAN integration – 22%
- Symantec Windows ASL – 24%
- Microsoft DPM integration – 27%
9:02 am – Audience questions focusing on dedupe – lots of interest and curiosity! More drill-downs will take place in individual sessions later today.