HP StorageWorks 4400 Enterprise Virtual Array (EVA4400) Performance

End-to-end performance numbers reflect typical data and code paths used in a customer situation. These numbers are an excellent choice for capacity planning when obtained under two additional engineering constraints.
First, when evaluating performance under random workloads, HP recommends using uniform probability to access most, if not all, usable array capacity. This measurement is more accurate than “short-stroking (destroking),” a configuration that allows the disk to access only a small fraction of the drives’ capacity, thus enhancing performance numbers. Measuring performance by accessing all of an array’s capacity leads to sizing estimates that will not become obsolete over time as the customer consumes more of the array’s capacity.
The second constraint applies to the recognition from HP that response times are very important to customers and their business applications. HP publishes “end-to-end” throughputs at average response times of 30 milliseconds or less. Because cache numbers are not useful in this context, HP does not apply this constraint to cache-only measurements.
HP best practices recommend using end-to-end performance numbers for capacity planning because they are:

- Measurements that can be used for sizing in a real-life business condition

- Easy to verify at the customer site using a reasonable “black box approach” by eliminating bottlenecks outside the array and balancing the load across an array’s controllers and host ports

End-to-end performance numbers provide customers with the best data to determine which array configuration makes the best business, technical, and financial sense.
Cache performance numbers
Cache performance numbers register the largest throughput numbers possible with an array but do not reflect use under normal business conditions, diluting their usefulness for capacity planning.
Such cache-only workloads place a lighter load on an array’s internal buses than would exist at a customer site. Exercising only a part of an array’s data and code paths does not reflect actual response times or latencies that normally occur because the cache-only data path is shorter than the normal data path. Typically, these are the paths used when 100% of all I/O requests are satisfied from the array’s cache. Cache performance numbers are unsustainable in day-to-day operations, though exceptions will occur momentarily and show up as short-lived transients.
Cache serves as an intermediary between the servers and the disk drives. When viewed separately, cache performance numbers provide an artificial view of an array’s performance capabilities. What is most important about cache performance is the interaction of the cache with a disk array’s disk drives.
Because disk overload is a possible consequence of using cache performance numbers for capacity planning, HP best practices do not recommend this measurement.
EVA4400 performance summary
The HP EVA4400 offers real-world performance at a higher level than the EVA4100 and is more affordable. With the EVA4400, HP engineers aimed to lower the customer’s total cost of ownership (TCO) while delivering more capacity and better performance—all within the context of the existing EVA enterprise-class functionality.
The vast majority of customers use EVA arrays for two classes of workloads: large block sequential and small block random. The EVA4400 achieves improved sequential performance levels and at a reduced cost and complexity. With respect to random workloads, the performance of the EVA4400 scales with the number of its disk drives, allowing customers to choose configurations that best meet their needs.
The elevated levels of sequential performance provides customers improved data transfer rates for applications such as data warehousing, streaming video, high-performance technical computing, and backups and restores. Increased random workload performance translates into better throughputs and response times for applications such as file systems, transaction-orientated databases, and email.
The following two tables display the difference between the EVA4400 and its predecessor, the EVA4100. HP measured each array with the maximum number of drives supported (56 drives on the EVA4100 and 96 drives on the EVA4400). The random workload performance of the EVA4400 is approximately twice that because of the increased number of disk drives coupled with increased controller capability. The gain in internal controller bandwidth of the EVA4400 accounts for the marked improvement under the large block sequential workloads.