The IBM DS8000 family is the leader in the external storage market for mainframe environments; these reliable data systems have been deployed in business-critical environments for more than two decades, combining proven capabilities with technological innovation. Some of the most relevant aspects that make IBM DS8000 so unique are describe below:
The DS8000 is based in a three-layered “share everything” architecture.
- Host adapters (also known as front-end adapters) manage the I/O interfaces to other systems. They also manage the Fibre Channel path protocols for host I/Os and for replicating data to remote DS8000s.
- IBM Power-based controllers which are active-active Power servers positioned between the adapters which are the key elements of the performance, availability and efficiency delivered by this architecture.
- Device adapters (also known as back-end adapters) manage the internal storage devices. They also manage the SAS paths to drives, RAID protection, and drive sparing.
This architecture maximizes performance, availability and data resilience while reducing custom components and design complexity.
Tasks are balanced across all components, and in fact much of the capacity is offloaded to the adapter hosts, and therefore, if a node was to fail, the notion of losing 50% of the system is not correct. Customers may lose about 50% of the controller processing power and memory but all the host adapters have access to both nodes. Based on the workload and what they are doing, the performance impact will be minimum.
IBM Power processors inherited the firmware from the mainframe that handles error management. This eliminates the need of thousands of lines of code makes the microcode simpler, and simple means reliable. Error management in the firmware is also much faster than in other architectures.
The host and array PowerPC processors implement Application Specific Integrated Circuits (ASICS). This means that they are specifically designed for their mission. The ASICS enable massive parallelism of I/O between the cache and disk arrays. Unlike other arrays, the DS8000 allows for each disk in an array to directly process I/O’s in parallel with all the other disks in every array.
Optimized cache processing
Patented IBM algorithms dramatically improve cache processing. IBM DS8000 implements Adaptive Replacement Caching (ARC). Before ARC, the Least Recently Used (LRC) algorithm cleared cache of the oldest data to make room for new writes. But what if the oldest data was the data that benefitted from cache residency the most? ARC makes that determination. It clears the cache of the least-effective data, and as a result, IBM DS8000 cache behaves as though it was twice as large as it actually is. Along with ARC is more effective cache segmentation. DS8000 uses a 4k segment size. Other implementations use 16k or 64k segment sizes. So, for any data write that is larger than 4k there is cache waste with other arrays.
Finally, if you replicate data to remote sites, IBM DS8000 offers advantages as well. Compared to other implementations, DS8000 continuously transmits groups of consistent data to the secondary site as soon as possible after it is written in the primary, this implementation enables the fastest recovery time in the industry with 2 to 4 seconds RPO and < 60 seconds RTO at > 1,000 miles.
These three characteristics that have just been described, are just a sample of the many unique features provided by IBM DS8000. If you want to know more, please browse the following link https://www.ibm.com/downloads/cas/3Q6P3M0Q.