Far from sounding the death knell for VTLs, the API provides benefits for both VTL vendors and end users.
By Heidi Biggar
The dust has settled over Symantec’s decision to provide third-party vendor access to NetBackup’s application programming interface (API), and the virtual tape library (VTL) market appears no worse for wear despite forebodings in the press of its impending demise (this publication excluded).
Enterprise Strategy Group (ESG) research conducted days before Symantec’s OpenStorage API announcement last November showed strong and growing support for VTLs, and ESG does not expect the availability of the API to derail that growth. Of the survey respondents who currently back up to disk, 47% said they either already used VTLs or planned to implement the technology within 12 months (see figure).
The negativity around the OpenStorage API and VTLs stems from the simple fact that the API provides a common interface to all disk targets, regardless of type. NAS or VTL, it all appears to and is treated by NetBackup as disk, which means there is theoretically no need for a VTL interface in NetBackup environments.
For this reason, much of the initial coverage of the Symantec announcement focused on the negative impact of the API on the VTL market rather than on the potential efficiencies this type of native communication can bring to third-party VTL vendors and, importantly, the corresponding benefits it will result in for users.
By providing greater “visibility” into the way data is transmitted and integrated with NetBackup, the API should enable VTL vendors to spend a lot less time and money on reverse engineering to the NetBackup platform and a lot more time enhancing functionality, which ultimately will help VTL vendors differentiate themselves in the VTL market and stay competitive in the larger disk-based backup market, and users better meet evolving data-protection requirements with features such as disk virtualization, thin provisioning, high performance, clustering, capacity reduction (e.g., data de-duplication or single-instance storage), indexing and search, and replication.
Don’t discount VTLs
VTLs will play a key role in today’s data-protection environments—NetBackup or otherwise—for the foreseeable future. The following is a list of reasons why ESG believes that Symantec’s API does not spell the demise of VTLs:
- NetBackup isn’t the only backup/recovery application: While Symantec is a leader in backup-and-recovery software, other vendors have significant installed bases and they all will still need a virtual tape interface to disk.
- There are still use cases where multiple backup applications share disk resources: In those cases companies would likely prefer using a centralized solution, such as a VTL, that was designed to support multiple backup applications.
- Not everyone will upgrade to NetBackup 6.5 immediately: Many end users are reluctant to upgrade their backup software to the latest version because they do not want to disrupt their environments. As such, companies may hold off on upgrading to NetBackup 6.5 for some time.
- VTLs were developed and optimized for backup and recovery: In and of itself, emulating a tape library provides some value, but it cannot be the primary capability that VTL vendors promote. Certainly, by emulating a physical tape library on disk, there are little to no changes required in the backup software architecture or processes. The VTL behaves like tape, just faster. Backup-and-restore performance is significantly improved, and there is minimal overhead with provisioning disk. In addition to emulating a tape library, VTL systems bring significantly more capabilities, including clustering, thin-provisioning, and capacity optimization, and many VTL solutions offer mirroring or VTL-to-VTL replication for remote vaulting and disaster recovery. Again, ESG sees VTL solutions evolving into data-protection platforms over time.
- VTLs are turnkey: End users like fully integrated and quality-assured solutions. Using a generic disk array backup target means greater complexity. The VTL vendor has the responsibility to make sure that every backup application, driver, host bus adapter, etc., works with its solution. Users may have to pay a premium for VTLs compared to generic disk arrays. However, there are reasons that users implement VTLs: because they need the performance and scalability that VTLs provide, or they need a disk-based data-protection solution that is non-disruptive to existing backup processes and systems.
Ensuring the backup application’s catalog is kept up to date may also present a problem. If the media server is in control of the virtual-to-physical media copies, there is no problem. However, if the VTL controls the movement of data to tape, or replication between local and remote VTLs doesn’t involve the media server, then the backup application loses its ability to locate the copy for restore.As mentioned, the OpenStorage API benefits VTL manufacturers and other storage vendors that provide intelligent disk devices, as well as end users and Symantec.
For third-party VTL vendors, the API provides more information about the backup image, which is no longer just a “blob” of data received by the storage device. Vendors gain more visibility into the backup’s contents, which should allow them to leverage their own disk device’s features without reverse engineering. For example, vendors will be able to control what data is de-duplicated through NetBackup. How the capacity reduction is done (i.e., the third-party platform that is used) is transparent to NetBackup.
For end users, the API should translate into near-term tangible benefits as VTL vendors leverage the API for value—add purposes. VTL vendors should benefit from improved integration points and documentation for the NetBackup application—information that was previously unavailable to them. There are many potential benefits for users.
For example, backup administrators could leverage NetBackup to establish tiers for supported data-protection devices (disk and tape). Policies for retention and movement between tiers could be set within Net-Backup and executed by the disk device. Duplication, whether through disk-to-disk replication or cloning to tape, could be driven by NetBackup and, more importantly, replicas could be tracked in the NetBackup catalog.
For Symantec, the API puts the company back in the data-protection driver’s seat. Based on user-defined policies, NetBackup “controls” the data-protection environment, specifying what is targeted for protection and when actions (backup, replication, de-duplication, etc.) take place. Intelligent disk device vendors control how the process takes place. The API also enables Symantec to improve the management of shared storage (in conjunction with third-party products) as well as configuring and provisioning disk without having to re-architect the backup system.
It makes sense that Symantec would want to be in command and dictate how the disk vendors will interoperate with NetBackup. Doing so makes NetBackup more than a tape initiator; it potentially becomes a single restore interface and policy engine, centralizing administration for NetBackup-based environments.
However, the API has potentially significant benefits for VTL vendors, which in turn benefits end users. The OpenStorage API gives VTL vendors the opportunity to focus their R&D on feature/technology enhancement and not on reverse engineering to the NetBackup platform. VTL vendors that use the next few months wisely should find themselves in a much stronger position than they are today—and so, too, will VTL users. VTL vendors need to use this time to provide and/or enhance a wide range of features, such as capacity reduction, performance, and optimization.
Heidi Biggar is an analyst with the Enterprise Strategy Group (www.enterprisestrat egygroup.com).
OpenStorage API up-close
Symantec’s NetBackup OpenStorage API provides a common interface for all disk targets and what Symantec refers to as “intelligent disk devices,” including virtual tape libraries (VTLs) and disk-to-disk (D2D) appliances. The API enables these devices to integrate directly with NetBackup 6.5, which is due in the second quarter. The new release is also designed to integrate with Symantec’s own disk-based backup capabilities (e.g., PureDisk) down the road. In NetBackup 6.5, the NetBackup Media Server acts as the proxy between NetBackup and the disk target, treating the device connected to the Media Server as a native disk device rather than tape.
The OpenStorage API puts Symantec in a potentially better position to control data-protection processes in NetBackup environments and it has the added benefit of dealing with the pesky problem of NetBackup’s media catalog not being synchronized when the intelligent disk device initiates duplication to tape or a secondary location. From a third-party VTL vendor standpoint, the API provides much better “visibility” into the data being transmitted and integration with the NetBackup platform, which should allow VTL vendors to spend less time on basic backup-and-recovery integration issues and more time developing true value-add capabilities, such as disk virtualization, thin provisioning, high performance, clustering, capacity reduction (i.e., single instancing or data de-duplication), indexing and search, and replication.
Access to the OpenStorage API is available to third-party vendors belonging to the Symantec Technology Enabled Program (STEP). As of this writing, STEP vendors include Copan, Data Domain, Diligent, EMC, FalconStor Software, Network Appliance, Quantum, Sepaton, and Sun. The API is available to STEP partners today so that vendors can initiate the necessary integration work. Solutions leveraging the API should be available beginning in Q2 2007 when Symantec introduces NetBackup 6.5. The API will ship with every NetBackup 6.5 Media Server.