Working Principles
This section describes concepts and working principles of SmartPartition.
Concepts
SmartPartition isolates different sizes of cache resources for different services, ensuring QoS of mission-critical services. The cache capacity of an application is the cache resources that the application can use in a storage system. The cache capacity is the most important factor that affects the performance of a storage system. A larger cache capacity is translated into:
- Higher write combination rate, higher write hit ratio, and better disk access sequence for an application that delivers write I/Os.
- Higher read hit ratio for an application that delivers read I/Os.
- For an application that delivers sequential I/Os, its cache capacity does not need to be large but must meet I/O requirements.
- Higher access rate and better performance for an application that delivers random I/Os.
Cache resources are divided into read cache and write cache.
- Read cache improves the read hit ratio using the read prefetch mechanism.
- Write cache accelerates hosts' access to disks using methods like combination, hit, and sorting.
Storage systems allow you to set a dedicated read cache and write cache for each SmartPartition partition. Storage systems' cache partitions are classified into SmartPartition partitions and default partition.
- SmartPartition partitions are cache partitions set by users. They provide cache resources for applications in partitions.
- The default partition is a cache partition reserved by the system. It provides cache resources for system running and other applications that have no SmartPartition partitions.
Implementation
Figure 1-1 shows the implementation of SmartPartition.
SmartPartition partitions are created based on service LUNs or file systems. Each SmartPartition partition is exclusively accessed and does not interfere with other partitions. After you manually set cache capacity for a partition, SmartPartition will periodically analyze the number of I/Os processed by each partition to achieve an optimal configuration and ensure the QoS of mission-critical services.