If you are saturating your Fibre Channel (FC) network links to complete disk-to-tape data backup jobs in smaller and smaller timeframes, it may be time to consider installing a virtual tape library (VTL)
In this tutorial on virtual tape library technology, learn about buying and managing a VTL, VTLs and data deduplication, and virtual tape library vendors and products.
VIRTUAL TAPE LIBRARY TECHNOLOGY TUTORIAL TABLE OF CONTENTS
A virtual tape library is an enterprise data storage system that emulates physical tape drives and stores backup data on array-based, RAID-protected disk drives. By emulating a physical tape library, very few changes are usually required in the backup software architecture and processes. The virtual tape library works sequentially like tape and writes data to disk in tape format. Since VTLs emulate tape storage systems, organizations of all sizes can integrate disk-based backup into systems with tape-based libraries relatively pain free.
Larger enterprises may front tape-backup systems to gain the disk-based advantages, and integrate VTLs into disk-based offsite replication and disaster recovery (DR) systems. Branch offices, remote offices and departmental offices can also use VTLs to interface with any centralized tape-based backup systems. Jeff Boles, a senior analyst with the Taneja Group, said SMBs with tape-based backup systems can use VTLs to replace the tape without replacing the entire backup system. "[VTLs] certainly can be full-fledged replacements for tape if you still want to continue to work in a tape paradigm and manage backups in a tape-job format," Boles said.
The ability to shed duplicate data and exponentially increase the amount of data that can be stored is one of disk-based storage's biggest advantages over tape. A virtual tape library can deduplicate your data sets and either store the deduplicated data on tape, which means you would be able to store more data on each tape cartridge, or send the deduplicated data to an offsite location for disk-to-disk disaster recovery.
According to Boles, real-world duplication ratios hover around 15:1. Data deduplication is becoming a must-have feature in VTLs, according to many analysts.
Before buying a virtual tape library, first set your recovery time objectives (RTOs) and recovery point objectives (RPOs). How long can a system be down before revenue loss becomes a factor? How recent must your backup data be for a functional recovery?
Next, determine the backup system performance metrics you'll need to meet those objectives. Consider how much data you will regularly backup and how long of a backup window you will have. Then, shop VTL systems for speeds and feeds.
When buying a virtual tape library, it's important to look at your long-term needs. Can you add appliances in the future to spread out the workload, or can you repurpose the unit for use in branch and remote offices as you grow?
You should also consider when and how the proposed virtual tape library systems accomplish deduplication. While the deduplication speeds and feeds may be impressive, how available the data is while awaiting deduplication, and how often the VTL performs deduplication may hamper your recovery goals. The type of data your VTL will see is another important factor. If you are running more than one protection program, make sure your VTL can handle multiple data formats.
There are a number of things to consider regarding virtual tape library management. You have to install, configure and manage how your virtual tape library interacts with your backup software, your existing tape libraries, and your offsite replication system (if you are using replication). You'll have to implement and manage any deduplication features you install. Other things to consider include: Will you continue to use tape backup? Will you deduplicate your backup data and send it offsite or put it on the tape cartridges? All of these considerations can add to management complexity.
The good news is that, according to Boles, vendors are working to make VTL administration easier. Specifically, enterprise data storage and VTL vendors are making storage systems interact better with VTLs and recognize what's happening behind the VTLs so the data backup and recovery system can adapt.
Today, backup systems might send data to VTLs for disk-based storage, then send another stream offsite for disaster recovery purposes, and send another stream to tape for offsite archiving. That's a tremendous amount of load on the backup system.
Boles said soon the backup systems will recognize if the VTLs can copy and move data, and if so, the backup system may only send one data stream to the VTL, and let the VTL send data offsite for replication and copy the data to tape for archiving. Then the VTLs would be helping backup systems, not burdening them. Newer VTLs are also offering network-attached storage (NAS) support to interconnect with as many network protocols as possible.
Boles also said that some virtual tape libraries and backup systems are integrating the technologies together, such as Symantec Corp.'s OpenStorage program. Symantec partners that join the company's Symantec Technology Enabled Program (STEP) gain access to Symantec's OpenStorage API, and can write code to integrate their technologies with Symantec's Veritas NetBackup product. Other major VTL vendors are participating in Symantec's STEP program, including FalconStor Software Inc. and Data Domain Inc. (recently acquired by EMC Corp.).
There are a number of smaller companies as well as well-known storage system vendors that offer virtual tape library systems for the largest enterprises down to small- to midsized businesses (SMBs). The following are just a few of the many VTL options available.
EMC Corp. positions its LAN-based DL 1500 and 3000 at medium-sized businesses. Both products use 1 TB, 7200 RPM SATA disk drives. The DL 1500 has up to 36 TBs of usable capacity, includes six Gigabit Ethernet ports and two SAN ports, provides both NAS and VTL backup interfaces, and can move up to 720 GB per hour, the company says. The DL 3000 scales to 148 TB of usable capacity, includes six Gigabit Ethernet ports and four Fibre Channel ports, and moves data at up to 1.4 TB per hour.
The Quantum Corp. DXi3500 has up to 4.2 TB of useable capacity and 84 TB of retention capacity with a 20:1 deduplication ratio. It features both 10/100/1000 BaseT Ethernet, 2 Gbps Fibre Channel, or iSCSI host-to-appliance interfaces, up to 290 Mbps data transfer, and 1,600 virtual tape cartridges. The DXi3500 also provides NAS, VTL, or mixed presentation backup interfaces.
The Sepaton Inc. S2100-DS2 Express Virtual Tape Library is a SATA RAID-based storage appliance with two, 2 Gbps Fibre Channel ports that move data at up to 300 Mbps and has up to 7 TB of useable capacity, according to the company. It includes software compression, data deduplication, and remote replication capabilities, and can provide up to 192 virtual devices simultaneously and up to 2,048 virtual tape cartridges. Press materials also say the S2100-DS2 can be installed and backing up data in an existing tape environment within 30 minutes.
This was first published in August 2009