HP SmartCache technology

Technical white paper | Product, solution, or service
7
The Smart Array P421 controller disk configuration included the following:
Boot Volume: 2 - 72GB 10K 6G SAS drives configured in RAID 1
Database Logs: 4 - 300GB 10K 6G SAS drives configured in RAID 5
Database: 12 - 300GB (10K 6G SAS drives configured in RAID 1+0
HP SmartCache: 1 100GB SSD MLC 6G SSD drive configured in RAID 0. The HP SmartCache was configured in write-
through mode and assigned to cache the database volume
The database size was roughly 1.6TB during the two-hour run. Figure 5 shows the performance of the TPCe-like workload
both with HP SmartCache (red) and without HP SmartCache (blue). TPeS performance of the TPCe-like workload was
significantly better when using HP SmartCache. The SSD used as a cache was less than 7% the capacity of the database. The
cache-hit ratio of HP SmartCache was 84%, meaning that roughly five out of every six disk reads to the database volume
were serviced by the SSD versus the HDDs. A smaller amount of HP SmartCache would have yielded a smaller cache hit
percentage and lower performance gains.
Figure 5. TPCe-like workload performance with and without HP SmartCache.
Table 1 compares the database logical volume IO statistics without and with HP SmartCache. With the addition of the HP
SmartCache, performance results change dramatically simply by adding a 100-GB SSD.
Table 1. TPCe-like workload statistics
I/O Analysis
Without HP SmartCache
With HP SmartCache
Accesses 32,767,015 159,239,037 (almost 5x more)
Reads 29,114,040 140,159,417 (almost 5x more)
Writes 3,652,975 19,079,620 (5x more)
% reads 89 88
TPeS
Time in seconds