Like many in the storage industry, I keep wanting to declare tape is dead, or at least on its last legs, when it comes to data protection. I surely thought the triple whammy of emerging technologies like deduplication, asynchronous replication and removable disk cartridges were going to finally drive the nail through the heart of tape.
I couldn’t be more wrong. Last week, I had an insightful conversation with the president of a records management firm based in New Jersey which has, for years, serviced high-end customers like financial services and pharmaceutical firms in the Northeast. They also anticipated more of their clients replicating data to them and, for years, reserved space on their floor for disk for this purpose. Instead, they are installing more racks to hold tape and don’t foresee disk happening at all or not nearly to the scope they foresaw years ago.
They are finding that even though clients want or do replicate data, they also find users on the high end of the spectrum want control of the data in its useable format. Once replicated, users spin it off to tape and store it long term with them for long term archival and data recovery under potentially catastropic circumstances. They want the records management company to know as little about the data they are sending them as possible, and they want the data stored in a format inaccessable to anyone but the user. This is a trend that bears watching in the tape market for even as disk cartridge, deduplication and replication technologies take off, new reasons to keep data on tape are emerging.
FAN or File Area Network is the latest buzz word for file virtualization, coined by Brad O’Neill, senior analyst at The Taneja Group. (He has some big clients in this space.) In June 2006, O’Neill reported that a Taneja Group survey of global IT decision makers, found that 62% of respondents now identify “file management” as either “the top priority” or “one of their top priorities” requiring immediate attention in their data centers. Meanwhile, Tony Asaro, senior analyst at the Enterprise Strategy Group has just posted an interesting blog on the FAN market, or rather what he sees as the lack of one. So which is it? Where have all the FANs gone?
We received the following comment from a VAR based in Florida on our piece covering the newly proposed Fibre Channel over Ethernet (FCoE) standard:
The concept of FC over Ethernet has very limited value. According to this article, the FCoE consortium is targeting this at low-to-mid range servers, over 10 GbE, and as a convergence technology. While 10 GbE makes sense from the storage array to the switch, it makes little to no sense from the server to the switch for two reasons: first, low to mid range servers — where this is targeted, don’t have the I/O requirements to saturate 1 GbE much less 10 GbE, and their PCI busses would not be able to handle anywhere near 10GbE throughput (do the math), and second, the reason that FCP exists is not because of throughput, but deterministic response time, which is guaranteed by the FC protocol wheras the Ethernet protocol becomes more non-deterministic with high load. This lack of deterministic response time will not be fixed with FCoE.
For these reasons, FCoE to the server does not make sense on the low/mid (don’t need the throughput and couldn’t handle it anyway) or the high end (Ethernet lacks the predictable response time of FCP). So back to the question of FCoE over 10 GbE from the storage array to the switch — if the storage arrays were 10GbE capable, why not just use iSCSI, which is already supported and in wide use in the enterprise despite what some manufacturers’ marketing and media reports say? My personal opinion is that this is an effort on the part of manufacturers who are behind in iSCSI to change the game in an effort to compete, and provides little to no value to consumers.
We have the feeling this could develop into an interesting discussion in the industry over the next year or so as FC and iSCSI, originally at odds in the market, have increasingly been combined in tiered storage environments and multiprotocol systems. Still, combining the two protocols–especially in the same data stream–could become a thorny issue.
What do you think?
TechTarget’s networking reporter Andrew R. Hickey wrote an insightful piece on WAN optimization technologies embedded in Microsoft Windows Vista and Longhorn Server and how they could make separate WAN optimization boxes obsolete:
Vista and Longhorn contain redesigned TCP/IP stacks, quality of service (QoS) facilities, file systems, security systems, and WAN-friendly presentation layers for applications…TCP flow control and error recovery have been improved while remaining compatible with other TCP implementations… Microsoft has enhanced management control over QoS, meaning that network administrators might be willing to trust QoS markings from Windows machines. In addition, the native Windows file-system access protocol, CIFS, has been improved and will work with most existing applications without requiring program changes. Also, remote application delivery systems, like Windows Server Terminal Services or Citrix Presentation Server, will probably have their performance enhanced when applications are rewritten to use Vista’s Windows Presentation Foundation component.
Before you go chucking your Riverbed box out the window, though, there are a few caveats:
Vista’s security improvements interfere with some VPN clients, and certain security options could interfere with existing WAN performance or optimization products unless they’re disabled. Data-reduction compression done by external WAN optimization tools may still be very useful in some situations.
Enterprises should use caution and examine how compatible Microsoft’s technologies will be within their networks, according to Gartner.
“Windows Vista and Longhorn offer the promise of improved networking performance and security,” Gartner stated. “However, the scope and scale of the changes present significant security and compatibility risks. Most enterprises will delay large-scale deployments until after application compatibility has been verified, which Gartner expects to take 12 to 18 months. This will give networking components time to mature…. As a result, the benefits of the new Windows communications stack will not be broadly realized before 2009.”
For now, maybe try tossing a TV off the roof instead.
Unstructured data is… well, it’s unstructured. Unruly, even. Unbearable? Well, maybe that’s a bit too much, but it’s definitely hard to deal with. Word documents, email, images and MP3s, among other types of files, are crowding storage systems at an exponental rate and classifying this data can be very difficult. Analysts note that 85% of data in a typical enterprise is considered unstructured, and managing all of this data is a growing concern for many enterprises.
In this podcast, Pierre Dorion offers practical answers to the most common questions about unstructured data management he hears from storage pros today.
Download the Unstructured data management FAQ podcast.
It’s not quite a product announcement yet, but it’s worth noting here that the friendly folks at Mimosa say they’re going to use a new $17 million in Series C funding in part to launch two new products before the end of the year. The email archiving company, which claims 160-plus customers, says it’s going to add filesystems to the applications its NearPoint archive will support, which currently include emails, instant messages, and IP-based voicemails. The file archiving will be available this summer, according to TM Ravi, CEO. According to Ravi, SharePoint archiving will follow by the end of the calendar year.
Mimosa has amassed $34.5 million in VC funding since its founding in 2005, and had its first full year of revenue in 2006, but Ravi says he’s already thinking IPO. He was mum on the timing but said “the traditional venture-capital goal is to go public, and that’s our goal as well.” He added, slyly, “But if along the way someone makes an offer we can’t refuse…c’est la vie.”
On March 22, we posted a Q & A with Burton Group analyst Guy Creese about potential “gotchas” for companies considering Google Apps Premier Edition. Recently, we heard back from Google Enterprise product manager Rajen Sheth in response to Creese’s analysis.
Storage Soup: Guy Creese pointed out that Google partners with Postini for email archiving, but doesn’t offer archiving for documents and spreadsheets Do you plan to offer document archiving as well?
Sheth: I think we actually have a better story than what’s typically done with documents. Within an enterprise, documents are all over the place–they might be sitting on your laptop, they might be sitting on a file share somewhere, they might be in your email. What we provide is one central place that people can access [files]. If I want to search through all the documents I’ve ever made or gotten, I can just do a Google search on Docs and Spreadsheets.
In terms of bringing it out of Google’s repository, there are a couple of things that can be done. Number one, we allow the export of files in multiple file formats, and the second thing is we offer an XML API, which means you can access that data through XML and pull it out to a records management system or any other system, really that wants to leverage that data.
Storage Soup: Does that mean Google has no plans to get into that kind of records management or archival space?
Sheth: No. We consider ourselves a user collaboration package rather than a records management package, but we want to integrate with records management and email archiving systems people already have.
Storage Soup: What about the point Creese brings up about keyword search vs. some of the records management searches with other products–is that something Google might bring into Docs and Spreadsheets for retrieval?
Sheth: On a separate side of the business we offer clustered searches and taxonomies through Google Search Appliance. There has been some talk about offering that through these products, but what we’ve found with user studies, both with consumers and in the enterprise, that collaborative tagging of documents for future retrieval is more popular. People rarely do more than type in a couple of keywords–we’re working on a variety of things on our search capability where the front end remains the same, but on the back end we do the heavy lifting to figure out what people want.
Storage Soup: Creese said that “[Software as a service (SaaS)] companies will eventually get to the point where they can’t save everything. Even with storage prices dropping, as more and more corporations put their data into software as a service there’s going to be a tipping point coming, where either it starts to become expensive to save everything for the service and the service therefore raises its rates, or it’s just too difficult to find what’s there.” What’s Google’s response?
Sheth: I actually disagree with that. When we provide 10 GB [in Gmail inboxes] we’re doing that with the knowledge that we can serve that and serve that well at the price point that we’ve offered. There are things that are unique about how Google structures its data centers–we use commodity PC servers within our data centers and many of them to bring down storage costs, rather than large back end storage systems. So as a result of that I think we can serve storage for a much lower cost than most organziations are able to. So when we put out a package like this with a certain storage quota we do it with every intention that users will use that storage to the absolute hilt.
We’re already running millions upon millions of consumers on these services–in terms of scalability, we’ve already scaled the services to support many users. We’ve also already tested it with a large company–ourselves. Our users are very, very heavy email and document users, and we use our own systems, the same ones we have for our customers. We essentially battle-tested it to make sure that it could handle the load.
Storage Soup: Yahoo announced yesterday that they’re offering unlimited storage with their email. Is that where Gmail will go eventually?
Sheth: I think we’d go in a little bit different direction. What we’re focusing on is that when we give you a quota, we want to give you a variety of ways to use that quota, including large attachments. We want to offer a variety of ways to use that Gmail quota appropriately–provide very high quotas and provide more tools to use that storage.
Storage Soup: What’s Google’s response on Gmail’s limits on sending emails out to no more than 500 people per day?
Sheth: The issue that we have is that we want to protect our users from spammers. We don’t want spammers to use Gmail, whether it’s the regular edition or the premier edition. We also don’t want Gmail users receiving spam. That said, there is the need for the ability to deliver to multiple sets of users. A lot of organizations already have these mailing lists set up and can use them [with Gmail]. Over time, we see other parts of our product portfolio as candidates to help solve these types of problems. For example, we have a product called Google Groups, which provides the ability to have larger groups and group page, and it’s something people can use right now, and we consider it a good candidate to bring into the [Google Apps] platform down the road as well. We have groups with thousands and thousands of users.
Storage Soup: Creese also points out that Google lacks an equivalent to PowerPoint in your productivity suite.
Creese: My response to that is we’re definitely not trying to duplicate Microsoft Office. The way I would think of it is that Office is very well designed for individual productivity–an individual preparing something to present to a group of people. We’re focusing Google Docs and Spreadsheets on collaborative use case scenarios.
Here’s an interesting ancedote from a user requesting more storage at his company and likening the process to buying a car.
At least storage guys don’t wear those nasty shiny suits…
According to the New York Times, Yahoo! will offer subscribers unlimited email storage on their free webmail accounts starting in May. The company currently has a 1 GB mailbox limit; the move was attributed to “explosive growth in the size of attachments as people share ever more photos, music and videos via e-mail,” but we think it might also have something to do with rival Google making more and more noise in the email storage space.
As backup software vendors are discovering, being flexible is the name of the game when it comes to incorporating the management of some of today’s hottest storage technologies – CDP, data classification, data de-duplication or integration with VTLs – into their backup software.
I am also finding that when one tries to get updates from these companies, one needs to exercise some flexibility as well. Though I had indicated in my last blog entry that I planned to cover Symantec in this month’s blog, the two of us could not get our schedules in sync. So instead I spoke to CA and CommVault and plan to cover Symantec’s NetBackup in more detail next time – or so I hope.
This month I began by talking to Kelly Polanski, CommVault’s Director of Product Marketing, and during our conversation she gave me a statistic that set me back. She said that nearly 80% of CommVault Galaxy’s customer base already uses disk as their primary backup target – either in the form as a virtual tape library (VTL) or disk-as-disk.
This stat caught me off-guard since it contradicts what I have heard to date. For example, Bocada, an independent data protection management software product which reports on all major backup software products, recently told me they still typically see 75% of their customers using tape as their primary target for backups.
So, I did some checking to see if CommVault was like Superman in the backup software space or if other backup software vendors were seeing similar increases in their percentages of customers using disk as their primary target for backup.
Neither CA nor EMC could provide any definitive numbers as to what percentage of their customers were using disk as a primary target for backup though both know that their numbers are growing. Symantec had some numbers to share as they had recently completed an internal survey of 200 of their customers and found that 63% of them now use some form of disk-based protection.
On a side note – I do have to congratulate EMC on their strategy of boosting (inflating?) their numbers – devilish though it may be. EMC is finding more of their customers switching to disk, but they conveniently ship NetWorker with their VTLs. How much NetWorker functionality and licenses that EMC includes with each VTL I’m sure surely varies by how many billions of TBs of storage the customer buys. But, it should come as no surprise to anyone that backup to disk is escalating in new deployments of NetWorker in EMC customer sites.
Sarcasm aside, this rapidly rising rate of users backing up to disk numbers increases the urgency for backup software vendors to integrate the management of each of these different technologies. For as time-consuming as it is to log in to manage each CDP, replication and backup product, it becomes even more difficult to create a consistent set of policies across these products that ensure the level of data protection and recovery matches the application’s requirements.
Of course, the difficulty arises from the fact that each of these different products usually makes its own copies of data, has its own database and is driven by its own policy engine. From a global management perspective, this makes it almost impossible to achieve any consistent method of locating the right copy of data, applying policies centrally or really knowing where anything is.
Both CA and CommVault (I know, it took me a while to get here) address these issues but are taking different paths to do so. This month (March 2007), CA is releasing a service pack (SP) for their BrightStor ARCserve backup softwarethat will more closely tie together their ARCserve and WanSync replication software. This SP provides ARCserve with an interface into the WANSync product and allows ARCserve to backup copies of data that WANSync creates. While a step in the right direction, this is more of a patch job than anything really innovative.
CA’s longer term plan is much more intriguing, if they can pull it off. CA is leveraging its acquisition of MDY Group International and its enterprise records management software (soon to be named CA Records Manager) that they completed in June 2006 to lay the foundation for enterprise-wide policy management for any product database.
According to Kristi Perdue, CA’s Product Marketing Director for Information Management products, the CA Records Manager will provide users a centralized policy engine that they can apply to any vendor’s product data repository. Configured modularly with an open architecture, it permits organizations to use a common set of policies for any vendor’s replication or backup product. (I ought to be in marketing for CA, you think?)
Overall, not a bad idea, but unfortunately at this time it is still vaporware. Even though CA’s Perdue describes CA’s integration efforts as “very aggressive” in this area, I wouldn’t expect to see a product release from CA for at least another year.
Of the two, at least CommVault’s technology is real. All of their replication products – Galaxy, QuickRecovery, and ContinuousDataReplicator – use the same underlying database and share a common set of policies. It even extends to setting policies for performing data archiving and data migrations which is great – assuming you are using their product exclusively on all of your servers.
This is my main concern about CommVault, unless you are exclusively using CommVault’s product, you may still have to bring in something like CA’s Record Manager to manage CommVault along with all of your other backup software products. But, whether that is a flaw in CommVault’s product design or a larger indicator of how enterprises run their businesses or let their businesses run them is a topic for another day.