|
A new approach to data compression
Compression is as old as personal computing, but it is used today mainly for backup and second tier storage. Many attempts to address the challenges of primary storage compression have failed because of degraded performance and lack of transparency.
| Compression
- the Storwize Way |
Because
compression of primary storage data has been difficult to
address, most efforts in recent years have been directed at
compression during remote transport. This type of compression
saves bandwidth but does nothing to reduce stored data volumes.
Attempts to compress primary storage data interfered with
user workflow and required running software on both storage
and end-user devices. These solutions were based on cumbersome
sequential access (decompress, update, recompress), work at
file-level granularity, or both. They also wasted client resources
and introduced unacceptable latencies. Storwize technology
processes in real-time only the data required for a particular
user or application, and fully supports every type of file
access. The Storwize method offloads and centralizes compression
in a powerful, dedicated appliance, eliminating performance
degradation and in most cases actually improving it.
| |
Full
transparency for |
| |
Client
OS
Applications
Storage unit
|
| |
Real-time
random access to compressed data |
| |
Read from any location in the
compressed data
Modify any location in the compressed
data |
| |
Storage
performance remains the same or improves |
| |
No CPU or memory penalty for clients or storage
(compression process is performed in a sole-purpose appliance,
designed and optimized for this specific task)
Smaller datagram size on storage subsystem allows to reduce
storage workload and frees potential bottlenecks
(disks CPU, cache mechanisms and more
|
| |
Complementary
solution |
| |
Storwize
is complementary to other ILM/storage management/TCO reduction
solutions
The appliance is vendor disk
subsystem agnostic
|
|
 |
|