Performance testing at 2020-12

Owned by Vadim Reutskiy
Last updated: Jan 18, 2021
1 min read

Test runs

Test run I

Test run parameterValue
Test targetHyperledger Iroha 1.2.0
Test start time2020-12-TBD
Test end time2020-12-TBD
Test length in minutes~4320 (3 days)
Request data

1) Ordered batch 1 (CreateDomain, CreateAsset, CreateAccount, AddAssetQuantity, TransferAsset)

2) Ordered batch 2 (GrantPermission, SetAccountDetail)

3) Ordered batch 3 (SetAccountDetail)

Network configurationConfiguration B (4 nodes, 2 hosts, 5GB memory per node)
Iroha configIroha config B (500 max rounds delay, 1000 max proposal size)

Results

Trial runs were made and the results obtained the following data:

Metric typeResulting value
Requests sent179 231
Requests per second (TPS)4.6
Amount of blocks generated3092 (question)
Block store size per node3092
WSV size per nodeDidn't run in WSV mode


Conclusions

The current performance of the Iroha network is much lower than in previous releases. Hence, the team is concentrated now on finding the reason and making a patch that will make the performance at least not lower than before.

Test run II

Test run parameterValue
Test targetHyperledger Iroha 1.2.0
Test start time2020-12-TBD
Test end time2020-12-TBD
Test length in minutes~4320 (3 days)
Request data

1) Ordered batch 4 (CreateAccount, SetAccountQuorum, AddSignatory, AddAssetQuantity)

2) Ordered batch 5 (TransferAsset)

3) Ordered batch 6 (GetPendingTransactions)

Network configurationConfiguration B (4 nodes, 2 hosts, 5GB memory per node)
Iroha configIroha config B (500 max rounds delay, 1000 max proposal size)

Results

Metric typeResulting value
Requests sentTBD
Requests per second (TPS)TBD
Amount of blocks generatedTBD
Block store size per nodeTBD
WSV size per nodeTBD