quorum IBFT benchmarking

... the other consensus algorithm that is shipped with quorum.

Also see general infos in previous document (which documents the initial raft consensus benchmarking).

tweaking and benchmarking: environment used in this first part, mempool settings, faster blockperiod, higher gas Limit
example runs
7nodes - istanbul.blockperiod 1sec, gasLimit 20 million
crux docker 4nodes
Direct RPC call instead of web3 call: result > 400 TPS but only for the first 14k tx
on Amazon AWS
run 11 - purpose: new diagrams
how to further increase the TPS?
issues

tweaking and benchmarking

environment used in this first part

jpmorganchase/quorum-examples is

vagrant virtualbox
7 nodes
using - see bootstrap.sh L33 and L11:
- jpmorganchase/quorum and
- jpmorganchase/constellation

I have been trying to raise the TPS:

mempool settings

20,000 transactions cannot be propagated without changing the txpool settings (see quorum issue 467):

--txpool.globalslots 20000 
--txpool.globalqueue 20000

add to ARGS in 7nodes/istanbul-start.sh

ARGS="--txpool.globalslots 20000 --txpool.globalqueue 20000 ...

faster blockperiod

the default 7nodes example uses a 5 seconds istanbul.blockperiod, which can be changed with

--istanbul.blockperiod 1

add to ARGS in 7nodes/istanbul-start.sh

ARGS="... --istanbul.blockperiod 1 ...

higher gas Limit

in default settings, but high hammering rate ... every block turns out to be full, so I also tried to raise the gaslimit (from the default 4.7 million):

change istanbul-genesis.json to

--> 20 million gas ("gasLimit": "0x1312D00")

However, the effect is simply that then the blocktime is going up. So raising the gasLimit does not result in higher TPS yet:

10 mio gasLimit
20 mio gasLimit

both saturate around ~200 TPS.

example runs

7nodes - istanbul.blockperiod 1sec, gasLimit 20 million

./tps.py 

versions: web3 4.3.0, py-solc: 2.1.0, solc 0.4.23+commit.124ca40d.Linux.gpp, testrpc 1.3.4, python 3.5.3 (default, Jan 19 2017, 14:11:04) [GCC 6.3.0 20170118]
web3 connection established, blockNumber = 21, node version string =  Geth/v1.7.2-stable-159d813f/linux-amd64/go1.9.3
first account of node is 0xcA843569e3427144cEad5e4d5999a3D0cCF92B8e, balance is 1000000000 Ether
nodeName: Quorum, nodeType: Geth, consensus: istanbul, chainName: ???

Block  21  - waiting for something to happen
(filedate 1534170380) last contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34
(filedate 1534171683) new contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34

starting timer, at block 42 which has  1  transactions; at timecode 131887.871188314
block 42 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX    1 /  1.0 s =   1.0 TPS_average
block 43 | new #TX  88 / 1000 ms =  88.0 TPS_current | total: #TX   89 /  2.6 s =  33.8 TPS_average
block 44 | new #TX 277 / 1000 ms = 277.0 TPS_current | total: #TX  366 /  4.4 s =  83.6 TPS_average
block 45 | new #TX 527 / 1000 ms = 527.0 TPS_current | total: #TX  893 /  7.1 s = 126.6 TPS_average
block 46 | new #TX 600 / 3000 ms = 200.0 TPS_current | total: #TX 1493 / 11.0 s = 135.2 TPS_average
block 47 | new #TX 731 / 3000 ms = 243.7 TPS_current | total: #TX 2224 / 11.7 s = 190.5 TPS_average
block 48 | new #TX 755 / 4000 ms = 188.8 TPS_current | total: #TX 2979 / 18.5 s = 161.3 TPS_average
block 49 | new #TX 756 / 4000 ms = 189.0 TPS_current | total: #TX 3735 / 21.9 s = 170.6 TPS_average
block 50 | new #TX 756 / 4000 ms = 189.0 TPS_current | total: #TX 4491 / 26.4 s = 170.1 TPS_average
block 51 | new #TX 756 / 3000 ms = 252.0 TPS_current | total: #TX 5247 / 30.2 s = 173.8 TPS_average
block 52 | new #TX 756 / 4000 ms = 189.0 TPS_current | total: #TX 6003 / 34.2 s = 175.4 TPS_average
block 53 | new #TX 756 / 4000 ms = 189.0 TPS_current | total: #TX 6759 / 37.6 s = 179.8 TPS_average
block 54 | new #TX 757 / 4000 ms = 189.2 TPS_current | total: #TX 7516 / 38.8 s = 193.5 TPS_average
block 55 | new #TX 757 / 5000 ms = 151.4 TPS_current | total: #TX 8273 / 46.6 s = 177.7 TPS_average
block 56 | new #TX 757 / 4000 ms = 189.2 TPS_current | total: #TX 9030 / 49.3 s = 183.3 TPS_average
block 57 | new #TX 757 / 4000 ms = 189.2 TPS_current | total: #TX 9787 / 55.0 s = 177.8 TPS_average
block 58 | new #TX 757 / 4000 ms = 189.2 TPS_current | total: #TX 10544 / 59.1 s = 178.4 TPS_average
block 59 | new #TX 757 / 4000 ms = 189.2 TPS_current | total: #TX 11301 / 62.7 s = 180.2 TPS_average
block 60 | new #TX 758 / 5000 ms = 151.6 TPS_current | total: #TX 12059 / 67.4 s = 178.9 TPS_average
block 61 | new #TX 758 / 4000 ms = 189.5 TPS_current | total: #TX 12817 / 68.2 s = 187.9 TPS_average
block 62 | new #TX 758 / 4000 ms = 189.5 TPS_current | total: #TX 13575 / 75.8 s = 179.1 TPS_average
block 63 | new #TX 758 / 4000 ms = 189.5 TPS_current | total: #TX 14333 / 80.4 s = 178.3 TPS_average
block 64 | new #TX 758 / 4000 ms = 189.5 TPS_current | total: #TX 15091 / 83.6 s = 180.4 TPS_average
block 65 | new #TX 759 / 4000 ms = 189.8 TPS_current | total: #TX 15850 / 87.9 s = 180.4 TPS_average
block 66 | new #TX 759 / 4000 ms = 189.8 TPS_current | total: #TX 16609 / 91.5 s = 181.5 TPS_average
block 67 | new #TX 759 / 5000 ms = 151.8 TPS_current | total: #TX 17368 / 93.3 s = 186.1 TPS_average
block 68 | new #TX 759 / 1000 ms = 759.0 TPS_current | total: #TX 18127 / 93.7 s = 193.4 TPS_average
block 69 | new #TX 759 / 1000 ms = 759.0 TPS_current | total: #TX 18886 / 95.1 s = 198.6 TPS_average
block 70 | new #TX 760 / 1000 ms = 760.0 TPS_current | total: #TX 19646 / 96.2 s = 204.2 TPS_average
block 71 | new #TX 355 / 1000 ms = 355.0 TPS_current | total: #TX 20001 / 96.7 s = 206.9 TPS_average
block 72 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 97.1 s = 205.9 TPS_average
block 73 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 98.1 s = 203.8 TPS_average
block 74 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 99.2 s = 201.7 TPS_average
block 75 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 99.9 s = 200.2 TPS_average
block 76 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 100.9 s = 198.2 TPS_average
block 77 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 102.0 s = 196.2 TPS_average
block 78 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 103.0 s = 194.2 TPS_average
block 79 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 104.0 s = 192.3 TPS_average
block 80 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 105.0 s = 190.4 TPS_average
block 81 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 106.1 s = 188.6 TPS_average

https://github.com/drandreaskrueger/chainhammer/blob/master/reader/img/istanbul-1s-gas20mio_run2_tps-bt-bs-gas_blks40-85.png

crux docker 4nodes

different from above:

crux instead of constellation
consensys/quorum (instead of jpmc/quorum as above)
- Conor: "It uses Google’s protobuf for communication instead of a hand-coded serialisation library."
- me: "would that affect my benchmarking?"
- Conor: "it may be more efficient. but the data serialisation overhead is a small part of it - the crypto operations are far heavier"
docker instead of vagrant virtualbox
4 nodes instead of 7 nodes

For (txpool, blockspeed, gasLimit) parameter changes see my fork of their repo. For docker questions see crux issue 27.

N.B.: It can take many seconds until the chain actually starts moving, and without stuttering. So better wait a minute after starting with docker-compose -f docker-compose-local.yaml up --build.

log:

./tps.py 

versions: web3 4.3.0, py-solc: 2.1.0, solc 0.4.23+commit.124ca40d.Linux.gpp, testrpc 1.3.4, python 3.5.3 (default, Jan 19 2017, 14:11:04) [GCC 6.3.0 20170118]
web3 connection established, blockNumber = 17, node version string =  Geth/v1.7.2-stable-3f1817ea/linux-amd64/go1.10.1
first account of node is 0xcA843569e3427144cEad5e4d5999a3D0cCF92B8e, balance is 1000000000 Ether
nodeName: Quorum, nodeType: Geth, consensus: istanbul, chainName: ???

Block  17  - waiting for something to happen
(filedate 1534423209) last contract address: 0xc36C1118645ce456E0249562F44175953A076a7a
(filedate 1534423288) new contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34

starting timer, at block 25 which has  1  transactions; at timecode 16826.132793952
block 25 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX    1 /  0.9 s =   1.1 TPS_average
block 26 | new #TX 113 / 1000 ms = 113.0 TPS_current | total: #TX  114 /  2.1 s =  53.2 TPS_average
block 27 | new #TX 303 / 1000 ms = 303.0 TPS_current | total: #TX  417 /  3.4 s = 123.8 TPS_average
block 28 | new #TX 303 / 1000 ms = 303.0 TPS_current | total: #TX  720 /  4.0 s = 178.9 TPS_average
block 29 | new #TX 262 / 1000 ms = 262.0 TPS_current | total: #TX  982 /  5.3 s = 186.6 TPS_average
block 30 | new #TX 266 / 1000 ms = 266.0 TPS_current | total: #TX 1248 /  6.2 s = 201.6 TPS_average
block 31 | new #TX 287 / 1000 ms = 287.0 TPS_current | total: #TX 1535 /  7.1 s = 215.7 TPS_average
block 32 | new #TX 268 / 1000 ms = 268.0 TPS_current | total: #TX 1803 /  8.1 s = 223.7 TPS_average
block 33 | new #TX 286 / 1000 ms = 286.0 TPS_current | total: #TX 2089 /  9.3 s = 224.8 TPS_average
block 34 | new #TX 271 / 1000 ms = 271.0 TPS_current | total: #TX 2360 / 10.2 s = 230.9 TPS_average
block 35 | new #TX 285 / 1000 ms = 285.0 TPS_current | total: #TX 2645 / 11.2 s = 237.2 TPS_average
block 36 | new #TX 255 / 1000 ms = 255.0 TPS_current | total: #TX 2900 / 12.1 s = 239.5 TPS_average
block 37 | new #TX 277 / 1000 ms = 277.0 TPS_current | total: #TX 3177 / 13.1 s = 243.4 TPS_average
block 38 | new #TX 244 / 1000 ms = 244.0 TPS_current | total: #TX 3421 / 14.3 s = 239.6 TPS_average
block 39 | new #TX 264 / 1000 ms = 264.0 TPS_current | total: #TX 3685 / 15.2 s = 242.2 TPS_average
block 40 | new #TX 283 / 1000 ms = 283.0 TPS_current | total: #TX 3968 / 16.1 s = 245.8 TPS_average
block 41 | new #TX 276 / 1000 ms = 276.0 TPS_current | total: #TX 4244 / 17.1 s = 248.5 TPS_average
block 42 | new #TX 278 / 1000 ms = 278.0 TPS_current | total: #TX 4522 / 18.3 s = 246.9 TPS_average
block 43 | new #TX 272 / 1000 ms = 272.0 TPS_current | total: #TX 4794 / 19.2 s = 249.1 TPS_average
block 44 | new #TX 271 / 1000 ms = 271.0 TPS_current | total: #TX 5065 / 20.2 s = 251.1 TPS_average
block 45 | new #TX 264 / 1000 ms = 264.0 TPS_current | total: #TX 5329 / 21.1 s = 252.5 TPS_average
block 46 | new #TX 276 / 1000 ms = 276.0 TPS_current | total: #TX 5605 / 22.4 s = 250.8 TPS_average
block 47 | new #TX 290 / 1000 ms = 290.0 TPS_current | total: #TX 5895 / 23.3 s = 253.2 TPS_average
block 48 | new #TX 275 / 1000 ms = 275.0 TPS_current | total: #TX 6170 / 24.0 s = 257.6 TPS_average
block 49 | new #TX 271 / 1000 ms = 271.0 TPS_current | total: #TX 6441 / 25.2 s = 255.7 TPS_average
block 50 | new #TX 257 / 1000 ms = 257.0 TPS_current | total: #TX 6698 / 26.1 s = 256.5 TPS_average
block 51 | new #TX 281 / 1000 ms = 281.0 TPS_current | total: #TX 6979 / 27.1 s = 258.0 TPS_average
block 52 | new #TX 269 / 1000 ms = 269.0 TPS_current | total: #TX 7248 / 28.0 s = 258.8 TPS_average
block 53 | new #TX 273 / 1000 ms = 273.0 TPS_current | total: #TX 7521 / 29.2 s = 257.3 TPS_average
block 54 | new #TX 287 / 1000 ms = 287.0 TPS_current | total: #TX 7808 / 30.2 s = 258.9 TPS_average
block 55 | new #TX 270 / 1000 ms = 270.0 TPS_current | total: #TX 8078 / 31.1 s = 259.8 TPS_average
block 56 | new #TX 269 / 1000 ms = 269.0 TPS_current | total: #TX 8347 / 32.0 s = 260.6 TPS_average
block 57 | new #TX 273 / 1000 ms = 273.0 TPS_current | total: #TX 8620 / 33.3 s = 259.2 TPS_average
block 58 | new #TX 267 / 1000 ms = 267.0 TPS_current | total: #TX 8887 / 34.2 s = 259.9 TPS_average
block 59 | new #TX 273 / 1000 ms = 273.0 TPS_current | total: #TX 9160 / 35.1 s = 260.8 TPS_average
block 60 | new #TX 285 / 1000 ms = 285.0 TPS_current | total: #TX 9445 / 36.1 s = 261.9 TPS_average
block 61 | new #TX 271 / 1000 ms = 271.0 TPS_current | total: #TX 9716 / 37.0 s = 262.5 TPS_average
block 62 | new #TX 272 / 1000 ms = 272.0 TPS_current | total: #TX 9988 / 38.3 s = 261.1 TPS_average
block 63 | new #TX 280 / 1000 ms = 280.0 TPS_current | total: #TX 10268 / 39.2 s = 262.0 TPS_average
block 64 | new #TX 264 / 1000 ms = 264.0 TPS_current | total: #TX 10532 / 40.1 s = 262.5 TPS_average
block 65 | new #TX 252 / 1000 ms = 252.0 TPS_current | total: #TX 10784 / 41.1 s = 262.7 TPS_average
block 66 | new #TX 290 / 1000 ms = 290.0 TPS_current | total: #TX 11074 / 42.3 s = 261.8 TPS_average
block 67 | new #TX 298 / 1000 ms = 298.0 TPS_current | total: #TX 11372 / 43.2 s = 263.1 TPS_average
block 68 | new #TX 252 / 1000 ms = 252.0 TPS_current | total: #TX 11624 / 44.2 s = 263.2 TPS_average
block 69 | new #TX 280 / 1000 ms = 280.0 TPS_current | total: #TX 11904 / 45.1 s = 263.9 TPS_average
block 70 | new #TX 281 / 1000 ms = 281.0 TPS_current | total: #TX 12185 / 46.3 s = 263.0 TPS_average
block 71 | new #TX 287 / 1000 ms = 287.0 TPS_current | total: #TX 12472 / 47.3 s = 263.8 TPS_average
block 72 | new #TX 237 / 1000 ms = 237.0 TPS_current | total: #TX 12709 / 48.2 s = 263.7 TPS_average
block 73 | new #TX 276 / 1000 ms = 276.0 TPS_current | total: #TX 12985 / 49.1 s = 264.3 TPS_average
block 74 | new #TX 286 / 1000 ms = 286.0 TPS_current | total: #TX 13271 / 50.1 s = 265.1 TPS_average
block 75 | new #TX 294 / 1000 ms = 294.0 TPS_current | total: #TX 13565 / 51.3 s = 264.4 TPS_average
block 76 | new #TX 250 / 1000 ms = 250.0 TPS_current | total: #TX 13815 / 51.9 s = 265.9 TPS_average
block 77 | new #TX 276 / 1000 ms = 276.0 TPS_current | total: #TX 14091 / 53.2 s = 265.0 TPS_average
block 78 | new #TX 282 / 1000 ms = 282.0 TPS_current | total: #TX 14373 / 54.1 s = 265.6 TPS_average
block 79 | new #TX 275 / 1000 ms = 275.0 TPS_current | total: #TX 14648 / 55.0 s = 266.1 TPS_average
block 80 | new #TX 256 / 1000 ms = 256.0 TPS_current | total: #TX 14904 / 56.0 s = 266.1 TPS_average
block 81 | new #TX 283 / 1000 ms = 283.0 TPS_current | total: #TX 15187 / 57.2 s = 265.3 TPS_average
block 82 | new #TX 287 / 1000 ms = 287.0 TPS_current | total: #TX 15474 / 58.2 s = 266.0 TPS_average
block 83 | new #TX 270 / 1000 ms = 270.0 TPS_current | total: #TX 15744 / 59.1 s = 266.4 TPS_average
block 84 | new #TX 277 / 1000 ms = 277.0 TPS_current | total: #TX 16021 / 60.1 s = 266.8 TPS_average
block 85 | new #TX 270 / 1000 ms = 270.0 TPS_current | total: #TX 16291 / 61.3 s = 265.8 TPS_average
block 86 | new #TX 256 / 1000 ms = 256.0 TPS_current | total: #TX 16547 / 62.2 s = 266.0 TPS_average
block 87 | new #TX 281 / 1000 ms = 281.0 TPS_current | total: #TX 16828 / 63.1 s = 266.5 TPS_average
block 88 | new #TX 278 / 1000 ms = 278.0 TPS_current | total: #TX 17106 / 64.1 s = 266.9 TPS_average
block 89 | new #TX 275 / 1000 ms = 275.0 TPS_current | total: #TX 17381 / 65.1 s = 267.1 TPS_average
block 90 | new #TX 258 / 1000 ms = 258.0 TPS_current | total: #TX 17639 / 66.0 s = 267.1 TPS_average
block 91 | new #TX 283 / 1000 ms = 283.0 TPS_current | total: #TX 17922 / 67.3 s = 266.4 TPS_average
block 92 | new #TX 266 / 1000 ms = 266.0 TPS_current | total: #TX 18188 / 68.2 s = 266.7 TPS_average
block 93 | new #TX 287 / 1000 ms = 287.0 TPS_current | total: #TX 18475 / 69.1 s = 267.2 TPS_average
block 94 | new #TX 264 / 1000 ms = 264.0 TPS_current | total: #TX 18739 / 70.1 s = 267.4 TPS_average
block 95 | new #TX 288 / 1000 ms = 288.0 TPS_current | total: #TX 19027 / 71.3 s = 266.9 TPS_average
block 96 | new #TX 264 / 1000 ms = 264.0 TPS_current | total: #TX 19291 / 71.9 s = 268.1 TPS_average
block 97 | new #TX 274 / 1000 ms = 274.0 TPS_current | total: #TX 19565 / 73.2 s = 267.3 TPS_average
block 98 | new #TX 230 / 1000 ms = 230.0 TPS_current | total: #TX 19795 / 74.2 s = 266.9 TPS_average
block 99 | new #TX 206 / 1000 ms = 206.0 TPS_current | total: #TX 20001 / 75.2 s = 266.1 TPS_average
block 100 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 76.1 s = 262.8 TPS_average
block 101 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 77.1 s = 259.6 TPS_average
block 102 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 78.0 s = 256.4 TPS_average
block 103 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 78.9 s = 253.4 TPS_average
block 104 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 80.2 s = 249.5 TPS_average
block 105 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 81.1 s = 246.6 TPS_average
block 106 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 82.0 s = 243.8 TPS_average
block 107 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 83.0 s = 241.0 TPS_average
block 108 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 84.2 s = 237.5 TPS_average
block 109 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 85.2 s = 234.8 TPS_average
block 110 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 86.1 s = 232.3 TPS_average
block 111 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 87.1 s = 229.7 TPS_average
block 112 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 88.0 s = 227.3 TPS_average
block 113 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 88.9 s = 224.9 TPS_average
block 114 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 90.2 s = 221.8 TPS_average
block 115 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 91.1 s = 219.5 TPS_average

https://github.com/drandreaskrueger/chainhammer/blob/master/reader/img/istanbul-crux-docker-1s-gas20mio_tps-bt-bs-gas_blks26-111.png

until now the best result - around

273 TPS.

No blocktime increase (as before with constellation 7 nodes) but a steady 1 second blocktime. Blocks not are full in terms of gas, so: how to further increase the TPS?

Direct RPC call instead of web3 call

With this config switch change, all those 20k transactions are not each sent through the standard web3 call

contract.functions.set( x=arg ).transact(txParameters)

but the transaction is manually compiled, and sent via a POST call directly to the RPC API of the node:

requests.post(RPCaddress, json=payload, headers=headers)

(see full code here).

Quite surprising, but web3 seems to seriously harm the overall execution performance!

Now also a different number of multithreading workers is optimal, see run 8.

run	workers	TPS_average peak	TPS_average at end
4	5	434	356
3	10	439	364
7	12	442	363
8	13	447	369
9	14	437	363
5	15	435	362
6	23	428	362
2	100	437	369

run 8 --> 13 workers seem to result in the highest TPS rate.

log of best run 8:

./tps.py 

versions: web3 4.3.0, py-solc: 2.1.0, solc 0.4.23+commit.124ca40d.Linux.gpp, testrpc 1.3.4, python 3.5.3 (default, Jan 19 2017, 14:11:04) [GCC 6.3.0 20170118]
web3 connection established, blockNumber = 17, node version string =  Geth/v1.7.2-stable-3f1817ea/linux-amd64/go1.10.1
first account of node is 0xcA843569e3427144cEad5e4d5999a3D0cCF92B8e, balance is 1000000000 Ether
nodeName: Quorum, nodeType: Geth, consensus: istanbul, chainName: ???

Block  17  - waiting for something to happen
(filedate 1534430103) last contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34
(filedate 1534430266) new contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34

starting timer, at block 28 which has  1  transactions; at timecode 23804.15982982
block 28 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX    1 /  0.9 s =   1.1 TPS_average
block 29 | new #TX 213 / 1000 ms = 213.0 TPS_current | total: #TX  214 /  2.1 s =  99.7 TPS_average
block 30 | new #TX 563 / 1000 ms = 563.0 TPS_current | total: #TX  777 /  3.4 s = 229.3 TPS_average
block 31 | new #TX 509 / 1000 ms = 509.0 TPS_current | total: #TX 1286 /  4.4 s = 295.5 TPS_average
block 32 | new #TX 523 / 1000 ms = 523.0 TPS_current | total: #TX 1809 /  5.3 s = 341.3 TPS_average
block 33 | new #TX 393 / 1000 ms = 393.0 TPS_current | total: #TX 2202 /  5.9 s = 370.2 TPS_average
block 34 | new #TX 446 / 1000 ms = 446.0 TPS_current | total: #TX 2648 /  7.3 s = 365.0 TPS_average
block 35 | new #TX 467 / 1000 ms = 467.0 TPS_current | total: #TX 3115 /  8.2 s = 379.7 TPS_average
block 36 | new #TX 504 / 1000 ms = 504.0 TPS_current | total: #TX 3619 /  9.4 s = 383.1 TPS_average
block 37 | new #TX 464 / 1000 ms = 464.0 TPS_current | total: #TX 4083 / 10.1 s = 404.3 TPS_average
block 38 | new #TX 440 / 1000 ms = 440.0 TPS_current | total: #TX 4523 / 11.3 s = 398.8 TPS_average
block 39 | new #TX 487 / 1000 ms = 487.0 TPS_current | total: #TX 5010 / 12.3 s = 407.6 TPS_average
block 40 | new #TX 494 / 1000 ms = 494.0 TPS_current | total: #TX 5504 / 13.3 s = 414.5 TPS_average
block 41 | new #TX 459 / 1000 ms = 459.0 TPS_current | total: #TX 5963 / 14.0 s = 427.2 TPS_average
block 42 | new #TX 454 / 1000 ms = 454.0 TPS_current | total: #TX 6417 / 15.2 s = 421.6 TPS_average
block 43 | new #TX 435 / 1000 ms = 435.0 TPS_current | total: #TX 6852 / 16.5 s = 416.0 TPS_average
block 44 | new #TX 585 / 1000 ms = 585.0 TPS_current | total: #TX 7437 / 17.4 s = 426.8 TPS_average
block 45 | new #TX 451 / 1000 ms = 451.0 TPS_current | total: #TX 7888 / 18.1 s = 436.0 TPS_average
block 46 | new #TX 417 / 1000 ms = 417.0 TPS_current | total: #TX 8305 / 19.4 s = 427.5 TPS_average
block 47 | new #TX 473 / 1000 ms = 473.0 TPS_current | total: #TX 8778 / 20.4 s = 431.0 TPS_average
block 48 | new #TX 455 / 1000 ms = 455.0 TPS_current | total: #TX 9233 / 21.3 s = 433.2 TPS_average
block 49 | new #TX 460 / 1000 ms = 460.0 TPS_current | total: #TX 9693 / 22.0 s = 440.1 TPS_average
block 50 | new #TX 402 / 1000 ms = 402.0 TPS_current | total: #TX 10095 / 23.3 s = 433.6 TPS_average
block 51 | new #TX 504 / 1000 ms = 504.0 TPS_current | total: #TX 10599 / 24.3 s = 437.0 TPS_average
block 52 | new #TX 546 / 1000 ms = 546.0 TPS_current | total: #TX 11145 / 25.2 s = 442.0 TPS_average
block 53 | new #TX 386 / 1000 ms = 386.0 TPS_current | total: #TX 11531 / 26.2 s = 440.8 TPS_average
block 54 | new #TX 407 / 1000 ms = 407.0 TPS_current | total: #TX 11938 / 27.4 s = 435.7 TPS_average
block 55 | new #TX 475 / 1000 ms = 475.0 TPS_current | total: #TX 12413 / 28.3 s = 438.0 TPS_average
block 56 | new #TX 556 / 1000 ms = 556.0 TPS_current | total: #TX 12969 / 29.3 s = 442.6 TPS_average
block 57 | new #TX 408 / 1000 ms = 408.0 TPS_current | total: #TX 13377 / 29.9 s = 446.7 TPS_average
block 58 | new #TX 452 / 1000 ms = 452.0 TPS_current | total: #TX 13829 / 31.5 s = 438.8 TPS_average
block 59 | new #TX 402 / 1000 ms = 402.0 TPS_current | total: #TX 14231 / 32.2 s = 442.2 TPS_average
block 60 | new #TX 289 / 1000 ms = 289.0 TPS_current | total: #TX 14520 / 33.2 s = 437.8 TPS_average
block 61 | new #TX 185 / 1000 ms = 185.0 TPS_current | total: #TX 14705 / 34.1 s = 431.3 TPS_average
block 62 | new #TX 291 / 1000 ms = 291.0 TPS_current | total: #TX 14996 / 35.3 s = 424.4 TPS_average
block 63 | new #TX 298 / 1000 ms = 298.0 TPS_current | total: #TX 15294 / 36.3 s = 421.6 TPS_average
block 64 | new #TX 298 / 1000 ms = 298.0 TPS_current | total: #TX 15592 / 37.2 s = 418.9 TPS_average
block 65 | new #TX 256 / 1000 ms = 256.0 TPS_current | total: #TX 15848 / 38.2 s = 415.2 TPS_average
block 66 | new #TX 253 / 1000 ms = 253.0 TPS_current | total: #TX 16101 / 39.4 s = 408.6 TPS_average
block 67 | new #TX 268 / 1000 ms = 268.0 TPS_current | total: #TX 16369 / 40.4 s = 405.6 TPS_average
block 68 | new #TX 259 / 1000 ms = 259.0 TPS_current | total: #TX 16628 / 41.3 s = 402.6 TPS_average
block 69 | new #TX 294 / 1000 ms = 294.0 TPS_current | total: #TX 16922 / 42.0 s = 403.1 TPS_average
block 70 | new #TX 257 / 1000 ms = 257.0 TPS_current | total: #TX 17179 / 43.2 s = 397.5 TPS_average
block 71 | new #TX 283 / 1000 ms = 283.0 TPS_current | total: #TX 17462 / 44.2 s = 395.3 TPS_average
block 72 | new #TX 267 / 1000 ms = 267.0 TPS_current | total: #TX 17729 / 45.2 s = 392.6 TPS_average
block 73 | new #TX 260 / 1000 ms = 260.0 TPS_current | total: #TX 17989 / 46.1 s = 390.0 TPS_average
block 74 | new #TX 268 / 1000 ms = 268.0 TPS_current | total: #TX 18257 / 47.1 s = 387.4 TPS_average
block 75 | new #TX 281 / 1000 ms = 281.0 TPS_current | total: #TX 18538 / 48.1 s = 385.2 TPS_average
block 76 | new #TX 296 / 1000 ms = 296.0 TPS_current | total: #TX 18834 / 49.4 s = 381.3 TPS_average
block 77 | new #TX 235 / 1000 ms = 235.0 TPS_current | total: #TX 19069 / 50.0 s = 381.0 TPS_average
block 78 | new #TX 232 / 1000 ms = 232.0 TPS_current | total: #TX 19301 / 51.3 s = 376.2 TPS_average
block 79 | new #TX 281 / 1000 ms = 281.0 TPS_current | total: #TX 19582 / 52.3 s = 374.8 TPS_average
block 80 | new #TX 317 / 1000 ms = 317.0 TPS_current | total: #TX 19899 / 53.2 s = 373.7 TPS_average
block 81 | new #TX 102 / 1000 ms = 102.0 TPS_current | total: #TX 20001 / 54.2 s = 368.8 TPS_average
block 82 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 55.2 s = 362.5 TPS_average
block 83 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 56.1 s = 356.4 TPS_average
block 84 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 57.1 s = 350.5 TPS_average
block 85 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 58.0 s = 344.8 TPS_average
block 86 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 59.0 s = 339.3 TPS_average
block 87 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 60.2 s = 332.2 TPS_average
block 88 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 61.1 s = 327.1 TPS_average
block 89 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 62.1 s = 322.2 TPS_average
block 90 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 63.0 s = 317.3 TPS_average
block 91 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 63.9 s = 312.8 TPS_average
block 92 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 64.9 s = 308.3 TPS_average

result > 400 TPS but only for the first 14k tx

(diagram https://github.com/drandreaskrueger/chainhammer/blob/master/reader/img/istanbul-crux-docker-1s-gas20mio-RPC_run8_tps-bt-bs-gas_blks28-93.png)

Surprising acceleration when not using web3 ... not 273 TPS but

370 - 465 TPS !

Following detailed averages are calculated at the bottom of notebook reader/outdated/blocksDB_analyze_quorum-istanbul_RPC-call.ipynb

Averaging over blocks 40 to 80, we see averages of 370 - 393 TPS.

Looking at blocks 40 to 60 only, we see averages of ~465 TPS, but then it goes down considerably.
Looking at the later blocks 64 to 80 only, we see averages of ~270 TPS.

Unclear what exactly makes the rate collapse here, and after already approx 13000 - 14000 transactions. I have done several repetitions, they all look similar. Any ideas where this might come from?

(actually, same behavior as in geth Clique PoA)

on Amazon AWS

Ready-made AMI image - see reproduce.md#readymade-amazon-ami . Create such a machine, then:

cd ~
git clone  https://github.com/blk-io/crux blk-io_crux
cd ~/blk-io_crux/docker/quorum-crux

cp docker-compose.yaml docker-compose-local.yaml
nano docker-compose-local.yaml

edit for local build "to build the Docker images yourself", so that the top of docker-compose-local.yaml looks like this:

  node1: &quorum_crux_node
    # Pull image down from Docker Hub
    # image: blkio10/quorum-crux:v1.0.0
    # Uncomment the below, and comment out the above line to build the Docker images yourself
    image: blk.io/quorum/quorum-crux
    build:
      context: .
    container_name: quorum1

Edit the settings: Higher gas limit, larger txpool, blockperiod 1 second

sudo apt install jq
jq '.gasLimit = "0x1312D00"' istanbul-genesis.json > tmp && mv tmp istanbul-genesis.json

sed -i 's/PRIVATE_CONFIG/ARGS=$ARGS"--txpool.globalslots 20000 --txpool.globalqueue 20000 --istanbul.blockperiod 1 "\nPRIVATE_CONFIG/g' istanbul-start.sh

start docker build & run network:

docker-compose -f docker-compose-local.yaml up

Amazon AWS results

Exact same settings as above, running on a powerful c5.4xlarge machine:

./tps.py 

versions: web3 4.3.0, py-solc: 2.1.0, solc 0.4.24+commit.e67f0147.Linux.gpp, testrpc 1.3.4, python 3.5.3 (default, Jan 19 2017, 14:11:04) [GCC 6.3.0 20170118]
web3 connection established, blockNumber = 22, node version string =  Geth/v1.7.2-stable-3f1817ea/linux-amd64/go1.10.1
first account of node is 0xcA843569e3427144cEad5e4d5999a3D0cCF92B8e, balance is 1000000000 Ether
nodeName: Quorum, nodeType: Geth, consensus: istanbul, network: 1, chainName: ???, chainId: -1
Block  22  - waiting for something to happen
(filedate 1537882490) last contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34
(filedate 1537882504) new contract address: 0x9d13C6D3aFE1721BEef56B55D303B09E021E27ab

blocknumber_start_here = 27
starting timer, at block 27 which has  1  transactions; at timecode 179.104554098
block 27 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX    1 /  1.2 s =   0.8 TPS_average
block 28 | new #TX 266 / 1000 ms = 266.0 TPS_current | total: #TX  267 /  2.1 s = 124.9 TPS_average
block 29 | new #TX 624 / 1000 ms = 624.0 TPS_current | total: #TX  891 /  3.4 s = 264.1 TPS_average
block 30 | new #TX 639 / 1000 ms = 639.0 TPS_current | total: #TX 1530 /  4.3 s = 354.3 TPS_average
block 31 | new #TX 571 / 1000 ms = 571.0 TPS_current | total: #TX 2101 /  5.3 s = 399.3 TPS_average
block 32 | new #TX 559 / 1000 ms = 559.0 TPS_current | total: #TX 2660 /  6.2 s = 428.8 TPS_average
block 33 | new #TX 572 / 1000 ms = 572.0 TPS_current | total: #TX 3232 /  7.4 s = 434.1 TPS_average
block 34 | new #TX 571 / 1000 ms = 571.0 TPS_current | total: #TX 3803 /  8.4 s = 453.5 TPS_average
block 35 | new #TX 574 / 1000 ms = 574.0 TPS_current | total: #TX 4377 /  9.3 s = 469.3 TPS_average
block 36 | new #TX 572 / 1000 ms = 572.0 TPS_current | total: #TX 4949 / 10.3 s = 481.9 TPS_average
block 37 | new #TX 560 / 1000 ms = 560.0 TPS_current | total: #TX 5509 / 11.2 s = 491.4 TPS_average
block 38 | new #TX 570 / 1000 ms = 570.0 TPS_current | total: #TX 6079 / 12.5 s = 488.1 TPS_average
block 39 | new #TX 569 / 1000 ms = 569.0 TPS_current | total: #TX 6648 / 13.4 s = 496.2 TPS_average
block 40 | new #TX 575 / 1000 ms = 575.0 TPS_current | total: #TX 7223 / 14.0 s = 514.4 TPS_average
block 41 | new #TX 556 / 1000 ms = 556.0 TPS_current | total: #TX 7779 / 15.3 s = 509.0 TPS_average
block 42 | new #TX 563 / 1000 ms = 563.0 TPS_current | total: #TX 8342 / 16.2 s = 514.1 TPS_average
block 43 | new #TX 579 / 1000 ms = 579.0 TPS_current | total: #TX 8921 / 17.5 s = 510.7 TPS_average
block 44 | new #TX 591 / 1000 ms = 591.0 TPS_current | total: #TX 9512 / 18.1 s = 525.2 TPS_average
block 45 | new #TX 548 / 1000 ms = 548.0 TPS_current | total: #TX 10060 / 19.4 s = 519.8 TPS_average
block 46 | new #TX 566 / 1000 ms = 566.0 TPS_current | total: #TX 10626 / 20.3 s = 523.6 TPS_average
block 47 | new #TX 588 / 1000 ms = 588.0 TPS_current | total: #TX 11214 / 21.2 s = 528.0 TPS_average
block 48 | new #TX 552 / 1000 ms = 552.0 TPS_current | total: #TX 11766 / 22.2 s = 530.4 TPS_average
block 49 | new #TX 561 / 1000 ms = 561.0 TPS_current | total: #TX 12327 / 23.4 s = 526.2 TPS_average
block 50 | new #TX 571 / 1000 ms = 571.0 TPS_current | total: #TX 12898 / 24.4 s = 529.3 TPS_average
block 51 | new #TX 562 / 1000 ms = 562.0 TPS_current | total: #TX 13460 / 25.3 s = 531.8 TPS_average
block 52 | new #TX 573 / 1000 ms = 573.0 TPS_current | total: #TX 14033 / 26.3 s = 534.5 TPS_average
block 53 | new #TX 526 / 1000 ms = 526.0 TPS_current | total: #TX 14559 / 27.2 s = 535.3 TPS_average
block 54 | new #TX 503 / 1000 ms = 503.0 TPS_current | total: #TX 15062 / 28.4 s = 529.6 TPS_average
block 55 | new #TX 509 / 1000 ms = 509.0 TPS_current | total: #TX 15571 / 29.4 s = 529.9 TPS_average
block 56 | new #TX 535 / 1000 ms = 535.0 TPS_current | total: #TX 16106 / 30.0 s = 536.4 TPS_average
block 57 | new #TX 507 / 1000 ms = 507.0 TPS_current | total: #TX 16613 / 31.3 s = 531.3 TPS_average
block 58 | new #TX 501 / 1000 ms = 501.0 TPS_current | total: #TX 17114 / 32.2 s = 531.4 TPS_average
block 59 | new #TX 510 / 1000 ms = 510.0 TPS_current | total: #TX 17624 / 33.4 s = 526.9 TPS_average
block 60 | new #TX 481 / 1000 ms = 481.0 TPS_current | total: #TX 18105 / 34.1 s = 531.0 TPS_average
block 61 | new #TX 474 / 1000 ms = 474.0 TPS_current | total: #TX 18579 / 35.3 s = 525.7 TPS_average
block 62 | new #TX 524 / 1000 ms = 524.0 TPS_current | total: #TX 19103 / 36.3 s = 526.6 TPS_average
block 63 | new #TX 541 / 1000 ms = 541.0 TPS_current | total: #TX 19644 / 37.2 s = 527.9 TPS_average
block 64 | new #TX 357 / 1000 ms = 357.0 TPS_current | total: #TX 20001 / 38.1 s = 524.3 TPS_average
block 65 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 39.1 s = 512.0 TPS_average

fastest seen so far:

peak TPS_average is 536 TPS, final TPS_average is 524 TPS.

See reproduce.md#results for results on differently sized Amazon machines.

run 11

purpose: new diagrams
Amazon t2.xlarge AWS machine = 4 vCPUs, 16 GB RAM
chainhammer v38 with: ./send.py threaded2 23
quorum-crux commit f39db2345cf9d82e76d3905468e6e5ea1469b09d

./tps.py 
versions: web3 4.7.2, py-solc: 3.1.0, solc 0.4.24+commit.e67f0147.Linux.gpp, testrpc 1.3.5, python 3.5.3 (default, Sep 27 2018, 17:25:39) [GCC 6.3.0 20170516]
web3 connection established, blockNumber = 52, node version string =  Geth/v1.7.2-stable-3f1817ea/linux-amd64/go1.10.1
first account of node is 0xcA843569e3427144cEad5e4d5999a3D0cCF92B8e, balance is 1000000000 Ether
nodeName: Quorum, nodeType: Geth, consensus: istanbul, network: 1, chainName: ???, chainId: -1
Block  52  - waiting for something to happen
(filedate 1540076417) last contract address: 0x1932c48b2bF8102Ba33B4A6B545C32236e342f34
(filedate 1540076703) new contract address: 0x9d13C6D3aFE1721BEef56B55D303B09E021E27ab

blocknumber_start_here = 323
starting timer, at block 323 which has  1  transactions; at timecode 1536.297552707
block 323 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX    1 /  0.9 s =   1.1 TPS_average
block 324 | new #TX 103 / 1000 ms = 103.0 TPS_current | total: #TX  104 /  2.2 s =  47.8 TPS_average
block 325 | new #TX 487 / 1000 ms = 487.0 TPS_current | total: #TX  591 /  3.5 s = 170.2 TPS_average
block 326 | new #TX 554 / 1000 ms = 554.0 TPS_current | total: #TX 1145 /  4.5 s = 255.9 TPS_average
block 327 | new #TX 552 / 1000 ms = 552.0 TPS_current | total: #TX 1697 /  5.8 s = 292.6 TPS_average
block 328 | new #TX 365 / 1000 ms = 365.0 TPS_current | total: #TX 2062 /  6.2 s = 334.9 TPS_average
block 329 | new #TX 175 / 1000 ms = 175.0 TPS_current | total: #TX 2237 /  7.1 s = 313.4 TPS_average
block 330 | new #TX 412 / 1000 ms = 412.0 TPS_current | total: #TX 2649 /  8.4 s = 314.6 TPS_average
...
block 351 | new #TX 649 / 1000 ms = 649.0 TPS_current | total: #TX 11430 / 30.0 s = 381.3 TPS_average
block 352 | new #TX 383 / 1000 ms = 383.0 TPS_current | total: #TX 11813 / 30.4 s = 389.1 TPS_average
block 353 | new #TX 249 / 1000 ms = 249.0 TPS_current | total: #TX 12062 / 31.6 s = 381.5 TPS_average
block 354 | new #TX 231 / 1000 ms = 231.0 TPS_current | total: #TX 12293 / 32.2 s = 381.2 TPS_average
block 355 | new #TX 479 / 1000 ms = 479.0 TPS_current | total: #TX 12772 / 33.6 s = 380.0 TPS_average
block 356 | new #TX 489 / 1000 ms = 489.0 TPS_current | total: #TX 13261 / 34.3 s = 386.6 TPS_average
block 357 | new #TX 391 / 1000 ms = 391.0 TPS_current | total: #TX 13652 / 35.6 s = 383.0 TPS_average
block 358 | new #TX 292 / 1000 ms = 292.0 TPS_current | total: #TX 13944 / 36.4 s = 383.5 TPS_average
block 359 | new #TX 708 / 2000 ms = 354.0 TPS_current | total: #TX 14652 / 38.0 s = 385.8 TPS_average
block 361 | new #TX 172 / 1000 ms = 172.0 TPS_current | total: #TX 14824 / 39.3 s = 377.4 TPS_average
block 362 | new #TX 225 / 1000 ms = 225.0 TPS_current | total: #TX 15049 / 40.3 s = 373.7 TPS_average
block 363 | new #TX 322 / 1000 ms = 322.0 TPS_current | total: #TX 15371 / 41.5 s = 370.0 TPS_average
block 364 | new #TX 236 / 1000 ms = 236.0 TPS_current | total: #TX 15607 / 42.2 s = 369.8 TPS_average
block 365 | new #TX 183 / 1000 ms = 183.0 TPS_current | total: #TX 15790 / 43.5 s = 363.1 TPS_average
block 366 | new #TX 298 / 1000 ms = 298.0 TPS_current | total: #TX 16088 / 44.5 s = 361.8 TPS_average
block 367 | new #TX 318 / 1000 ms = 318.0 TPS_current | total: #TX 16406 / 45.5 s = 360.9 TPS_average
block 368 | new #TX 173 / 1000 ms = 173.0 TPS_current | total: #TX 16579 / 46.1 s = 359.3 TPS_average
block 369 | new #TX 292 / 1000 ms = 292.0 TPS_current | total: #TX 16871 / 47.7 s = 353.5 TPS_average
...
block 380 | new #TX 256 / 1000 ms = 256.0 TPS_current | total: #TX 19855 / 58.1 s = 341.8 TPS_average
block 381 | new #TX 146 / 1000 ms = 146.0 TPS_current | total: #TX 20001 / 59.0 s = 338.9 TPS_average
block 382 | new #TX   0 / 1000 ms =   0.0 TPS_current | total: #TX 20001 / 59.9 s = 333.7 TPS_average

Amazon t2.xlarge - TPS_average: peak 389.1 TPS, final 338.9 TPS

now with the new CLI tool, immediately create images:

./blocksDB_create.py temp.db
./blocksDB_diagramming.py temp.db quorum-crux-IBFT_t2xlarge 320 395

how to further increase the TPS?

any ideas? Please tell us --> quorum/issues/479

Actually, during the experiment my CPU is near 100% so I don't think this can be sped up much more.

Things to try:

send tx not all to 1 node :22001, but round-robin to :22001, :22002, :22003, :22004
compare dockerized 4nodes (blk-io/crux) with vagrant-virtualbox 4nodes branch, to get a grip on which virtualisation is faster. Environment is almost ready, just some last tweaks missing, see the comments below this one.

what else?

issues

IT#110 IBFT in parity?
P#9298 IBFT in parity?
IT#111 istanbul API not accessible from JSRE console?
Q#466 new release please
Q#467 txpool overflow issue [solved]
Q#478 please rename binary from geth to something else
Q#479 IBFT - 200 TPS max ? renamed to:
Q#479 Sudden drop in TPS around 14k transactions (Quorum IBFT)
GE#17447 Sudden drop in TPS after total 14k transactions
BC#27 4nodes quorum-example ?
Q#497 Please consider to catch up with geth
BC#27 4nodes quorum-examples - how to? to compare vagrant with docker.
BC#37 local docker build is failing: Service 'node1' failed to build
BC#39 ERROR: Service 'node1' failed to build: The command '/bin/sh -c chmod +x start.sh ...' returned a non-zero code: 2
BC#48 (feature request) quorum docker setup with 2-3 not 4 nodes (to run on small VPS)
BC#50 local build fails to start geth s
BC#51 (FR) dockerized with raft consensus not IBFT
BC#60 leveldb (missing)

TODO:

try this out, to compare vagrant with docker: BC#27

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

quorum-IBFT.md

quorum-IBFT.md

quorum IBFT benchmarking

TOC

tweaking and benchmarking

environment used in this first part

mempool settings

faster blockperiod

higher gas Limit

example runs

7nodes - istanbul.blockperiod 1sec, gasLimit 20 million

crux docker 4nodes

Direct RPC call instead of web3 call

result > 400 TPS but only for the first 14k tx

on Amazon AWS

Amazon AWS results

run 11

how to further increase the TPS?

issues

TODO:

Files

quorum-IBFT.md

Latest commit

History

quorum-IBFT.md

File metadata and controls

quorum IBFT benchmarking

TOC

tweaking and benchmarking

environment used in this first part

mempool settings

faster blockperiod

higher gas Limit

example runs

7nodes - istanbul.blockperiod 1sec, gasLimit 20 million

crux docker 4nodes

Direct RPC call instead of web3 call

result > 400 TPS but only for the first 14k tx

on Amazon AWS

Amazon AWS results

run 11

how to further increase the TPS?

issues

TODO: