buckyos/CYFS

Optimize the SN stress test client

Opened this issue · 1 comments

I have previously stress-tested the sn-miner service and obtained some performance data. (#164)
Now, we need to further understand the operation status of each client during the stress test, such as the response time of ping requests, the types of errors that occur, and the handling of exceptions, in order to understand the processing status of sn under extreme conditions.
So, I try improving the SN stress testing client tool to achieve the above features.

Regarding the stress test simulation, some key data of the test client need to be configured precisely.

1. The number of devices

Different DeviceIDs, generally one sn-client instance for one individual device

2. The ping interval of a single device

The default interval of bdt module is 15s, but the stress test tool can configure this interval, the default can be set to 1s

3. SN call interval

Device simulation sn call, also need to be able to configure the interval

4. The target device of sn call

To the same device call, or different device call (each call to choose a random device), should also be configurable

5. some exception simulation

For example, the wrong device public and private keys, call the target device is not online, etc., can be arranged as needed