Workload tests description

Edit online

The uperf tool was configured to provide 45 tests, separated into five workload types ranging from small transactional loads up to larger streaming loads. The five workload types are:

Highly transactional small data sizes (rr1c-1x1): This is a Request & Response (RR) test that sends a 1-byte request from the client and receives a 1-byte response from the server. This is the smallest possible transactional workload test and is smaller than most customer workloads. This test represents the RR overhead costs.
Highly transactional medium data sizes (rr1c-200x1000): Another Request & Response (RR) test that sends a 200-byte request from the client and receives a 1000-byte response from the server. This test should be representative of a typical user's interaction with a remote web site.
Large transactional data sizes (rr1c-200x30k): Another Request & Response (RR) test that sends a 200-byte request from the client and receives a 30,720-byte response from the server. This test should be representative of a typical user's interaction with a remote web site that contains large data items like pictures or music.
Streaming reads (str-readx30k): A Streaming (STR) test that transfers 30,720-byte datagrams from the client to the server. This test should emulate the behavior of workloads like backup or uploads.
Streaming writes (str-writex30k): Another Streaming (STR) test that transfers 30,720 bytes of data from the server to the client. This test emulates large downloads, streaming online content, etc.

User loads were simulated using varying connections counts (1, 10, 25, 50, 75, 100, 150, 200, 250) on each workload type. This resulted in a matrix of 45 unique workload tests.

Uperf requires two roles to perform tests, a client (uperf master) and server (uperf slave). In the comparison, the uperf client was assigned to the KVM guest, and the uperf server was assigned to the standalone LPAR.

For a detailed explanation of the uperf configuration and workload test compositions, refer to the Network workload section of the KVM Network Performance - Best Practices and Tuning Recommendations white paper.¹)

To learn how uperf was run, refer to the details provided in the Testing methodology section of the same paper.²

Note: This paper used an expanded set of connection/simulated user counts.

¹ Network workload, KVM Network Performance - Best Practices and Tuning Recommendations (https://www.ibm.com/docs/en/linux-on-systems?topic=recommendations-network-workload)

² Testing methodology, KVM Network Performance - Best Practices and Tuning Recommendations (https://www.ibm.com/docs/en/linux-on-systems?topic=recommendations-testing-methodology)