Author: Ian Miers 2012-10-04 16:31:11
Published on: 2012-10-04T16:31:11+00:00
For evaluating script evaluation performance, a good approximation of real-world traffic may not be enough, according to Joel Joonatan Kaartinen. Worst-case scenarios must also be benchmarked to avoid creating a DoS vulnerability. Ian Miers was primarily concerned with script evaluation performance and was trying to add new instruction types. To test how that affects performance, transactions with the new instructions need to be interspersed with existing ones. For accuracy, simulated traffic should at least approximate real-world traffic. Jeff Garzik suggested measuring specific aspects of blockchain performance by importing blocks via -loadblock=FILE. Other performance measurements such as how fast a block relays through the network cannot be easily measured. The potential for DoS attacks must be considered in security rather than performance when designing a test methodology.
Updated on: 2023-06-06T07:42:43.172129+00:00