Kaminario just launched Kaminario DataProtect, the newest implementation of the company's SPEAR software architecture using its K2 storage systems. While I was evaluating the content of their announcement, it became clear to me that the industry now expects storage systems to deliver new capabilities.
Storage systems, to meet today's requirements must both traditional rotating media, such as disk drives, and solid state storage made up of either dynamic random access memory (DRAM) and/or Flash storage.
The industry now expects intelligent storage systems and storage virtualization software to offer the following capabilities.
- Enterprise class high availability
- Redundant hardware making it possible for operations to continue if a component fails. This capability is usually referred to as having a "N+1" configuration.
- Hardware and software that are designed to create a "self-healing" environment that automatically and quickly moves data around so that operations continue without operator intervention.
- Fail over that does not disrupt current operations. This means storage systems must be designed using redundant hardware and heal itself automatically when a component fails,
- Enterprise class data protection
- Data protection features must be very fast to prevent data loss in real time
- Storage systems must support mirroring for data protection and striping for high performance.
- Snapshots of operational data must be available allowing back up to be done while systems continue to operate
- Provision for fast replication using asynchronous communications links providing for thorough disaster recovery
- Support for DRAM and Flash-based storage
- Hardware and software designed to support high throughput and low latency solid state storage devices
- Auto tiering designed to make best use of available storage devices
- Management tools making it possible for IT administrative staff to easily deal with a complex storage hierarchy
As production systems are ever more complex, it is becoming harder and harder to isolate and then fix application slow downs and failures. Kaminaro and others point out that many times, the issues can be tracked down to slow or badly configured storage systems or some other storage-related issue.
When I consider the new requirements for storage systems, only a very few suppliers can address all of the requirements.
For the most part, solid state storage has been added on as a caching mechanism, but not as part of the standard set of storage architectures traditional storage servers are designed to support. So, most of today's storage systems do not offer the range of performance, ease of management or availability offered by newer generation storage systems, such as those offered by Kaminario.
I expect that over time, all storage system suppliers will offer the same capabilities that Kaminario is offering today.
If your distributed workloads are not performing optimially, it might be better to examine more intelligent storage systems rather than simply throwing larger system configurations at the problem.