Choosing the Right Deduplication Solution for Your Organization

Client 1 Client 2 Client 3 Media server 1 Client 100 Media server 3 Media server 2 Media server 5 Media server 4 Media server 6 SAN or local storage 10 / 15K Figure 2. With software-based deduplication, avoiding logjams requires adding more media servers, which increases costs and complexity. To avoid logjams, customers often add media servers equipped with greater CPU and memory resources (see Figure 2). While this approach spreads out the extra load, it also obviously increases complexity and costs. Appliance-based deduplication Because purpose-built backup to disk appliances absorb deduplication and replication workloads, current backup environments can remain unchanged. Complexity is...
Client 1 Client 2 Client 3 Media server 1 Client 100 Media server 3 Media server 2 Media server 5 Media server 4 Media server 6 SAN or local storage 10 / 15K Figure 2. With software-based deduplication, avoiding logjams requires adding more media servers, which increases costs and complexity. To avoid logjams, customers often add media servers equipped with greater CPU and memory resources (see Figure 2). While this approach spreads out the extra load, it also obviously increases complexity and costs. Appliance-based deduplication Because purpose-built backup to disk appliances absorb deduplication and replication workloads, current backup environments can remain unchanged. Complexity is avoided because there is no need to beef up the number and horsepower of media servers (see Figure 4). 3 Purpose-built backup to disk appliances absorb deduplication and replication workloads, so the media server workload is limited to: • Processing client backup data • Creating tape copies from disk Figure 3. Appliance-based deduplication puts no extra load on the media server. With appliance-based deduplication, there is no need to beef up the number and horsepower of media servers. Client 1 Media server 1 Purpose-built deduplication appliances are built to simultaneously handle deduplication loads across multiple backup applications. Client 2 Client 3 Client 100 Media server 2 Purpose-built backup-to-disk appliance Figure 4. Appliance-based deduplication does not put additional load on media servers. What if the backup environment is already equipped with bigger and faster media servers with no logjams? Does this diminish the value of appliancebased deduplication? Definitely not. In fact, if desired, purpose-built backup to disk appliances can offload the deduplication processing from the appliance to a number of media servers. If enough media server resources become engaged, the deduplication process for backups can become faster than the appliance itself. Costs Vendors of backup software deduplication and compression solutions tout significant costs savings when compared to dedicated appliances, pointing out that existing investments in SAN, NAS and DAS storage systems can be utilized to support backup storage needs. However, these vendors fail to include in their calculations the tenfold increase in media server workloads with software deduplication solutions and the resulting significant cost of spreading this additional workload across a greater number of media servers with big CPUs and memory to avoid logjams. In addition, meeting service level agreements (SLAs) often requires fast — and therefore 4 Media server 3 expensive — solid state disk (SSD) drives be used for target repositories. These costs wash out the touted backup software deduplication and compression savings. Flexibility Software-based deduplication Each backup software application implements a proprietary approach to deduplication and compression. Thus, if an environment uses multiple backup software applications, two or more software-specific deduplication technologies are being implemented, which can make the environment more complex and difficult to manage. Appliance-based deduplication On the other hand, purpose-built deduplication appliances are not anchored to any specific backup software product — they are built to simultaneously handle deduplication loads across multiple backup applications. This capability is extremely valuable to IT environments that use multiple backup products. Moreover, as updates or changes to backup software occur in the future, the deduplication appliance can continue to be used, maximizing the value of the investment.
Read more...