5 keys to scalability with disk-based backup systems – Part 3 of 5
In part two of this blog series, we discussed how a grid-based system (instead of a single-controller/disk shelf model) allows backup performance to be maintained as your backup data grows. In part 3 of this 5 part series on scalability with disk backup systems, we discuss the issue of scalability with respect to how much investment in a system is required up-front in order to maintain backup performance over time.
- Does the system allow you to keep all of your backups on disk cost effectively?
- Does the system cause your backups to take longer as your backup data grows?
- Does the system require a larger up-front investment, or can you pay as you grow?
- Does the system require you to swap your system for a larger one (“forklift upgrade”) as your backup data grows, or can you preserve your initial investment when moving to a larger system?
- Does the core deduplication architecture lend itself to a more scalable model?
Key #3: Does the system require a larger up-front investment, or can you pay as you grow?
Related to whether backup performance can be maintained as the disk backup system expands, is the question of how much investment in a system is required up-front in order to maintain backup performance over time. As was mentioned in part 2 of this series, with a single-controller system, backup data growth is accommodated through the use of disk shelves by driving an increasing amount of data through a static amount of processing power, memory, and bandwidth. This results in backups taking longer as the amount of backup data increases.
One way to avoid having these larger backups exceed the backup window in a single-controller system is to power the single controller with very high-speed processors, and increased memory and bandwidth, from the beginning. That way, the increased amount of disk that is added later as the system expands can be accommodated.
While this can resolve the performance issues and keep backups within the desired backup window, this solution instead forces the customer to purchase a more powerful and more expensive disk backup system than they actually need at the time of purchase. Instead of purchasing the appropriate amount of processing power, memory, and bandwidth that is needed for the amount of backup data at the time of purchase, the customer is purchasing processing power, memory, and bandwidth to accommodate potential future data growth as well.
With a grid-based system, where each node in the grid contains the amount of CPU, memory, and bandwidth that is needed for the amount of disk on the node, the customer purchases additional CPU, memory, and bandwidth when they are needed, instead of up- front. In this way a grid-based system allows you to “pay as you grow”, instead of paying more up-front.
Use disk-based backup to drive more scalability to your IT operations. Download our free whitepaper, “Five Keys to Scalability and Disk-Based Backup Systems” for tips on how you can cost effectively and efficiently store all of your backup data.