| 14 Jul 2018 |
 komodobridge | [discord] <blackjok3r> http://cryptocartography.io/txscl_vis/ | 15:24:27 |
| komodobridge | [discord] <blackjok3r> https://youtu.be/pVEx1Hg_I7I | 15:28:11 |
| 15 Jul 2018 |
| komodobridge | [discord] <Mylo(KomodoPioneers)> That's awesome!! So when are we going to be doing some of these crazy tests? Will it be in the next:
a) 1 -3 days
b) 3-5 days
c) 7-10 days
d) ask tomorrow
😃 | 00:58:50 |
| komodobridge | [discord] <blackjok3r> That's not known at this stage, I await patchkez to return from vacation to help me with the last of the docker stuff, need to create a kubernetes template, and also Steve has to sort the last things out with AWS, so we have an account big enough to launch the full test. Some time in the next week I will run a full block chain test with real data that aims for 45k tx/s minimum to test out my docker containers running in kubernetes to verify everything works. Also need to get some notary nodes and explorer's for the first 64 chains ready. Maybe I can just run these in AWS as well, but in a different region. Not sure yet. | 01:23:10 |
| komodobridge | [discord] <smk762> Ive added a couple of thegns to the graph point tooltips - commas for number formatting, active chain count, and data state (live / simulated). | 03:36:10 |
| komodobridge | [discord] <smk762> https://i.imgur.com/vkulxpX.jpg | 03:36:11 |
| komodobridge | [discord] <smk762> Ive added a couple of things to the graph point tooltips - commas for number formatting, active chain count, and data state (live / simulated). | 03:37:03 |
| komodobridge | [discord] <smk762> Is there a plan to host the source data at a publicly available url for validation? I understand it is a lot of data... but I'd like to add a link for the curious. Maybe a link to contact details to someone which can provide it on request for a specific time period? | 04:08:04 |
| komodobridge | [discord] <blackjok3r> I think the blockchains will be saved in an AWS S3 data store, however running 8192 deamons to extract it would be difficult. I will work with dwy to adapt his script that dumps the data to a mysql DB. Maybe that can be put out to public for reads? | 04:35:08 |
| komodobridge | [discord] <smk762> Third party verification from an independent, trusted person/group is probably most valuable marketing-wise, but opening it to the public also shows good faith. Most wouldn't bother to check, but availability speaks volumes. Not sure what format is best. Database tables are easily populated with whatever, but a full copy of all chains blk.dat etc is closer to source, though a significant amount of data to review.
The scripts which are used to run the tests being open source also helps, at least for those that understand it. I don't think we'll have issues with burden of proof if the right people with no vested interest confirm it's valididty. | 05:01:03 |
| komodobridge | [discord] <smk762> sql dumps are small enough for public consumption and a good indicator of valididty. Official validation would likely want access to more | 05:02:15 |
| komodobridge | [discord] <smk762> I think that expanding the notes link on the vis site to be a pop-up onclick with links to sql and all relevant github repos isnt a bad idea | 05:03:36 |
| komodobridge | [discord] <smk762> Third party verification from an independent, trusted person/group is probably most valuable marketing-wise, but opening it to the public also shows good faith. Most wouldn't bother to check, but availability speaks volumes. Not sure what format is best. Database tables are easily populated with whatever, but a full copy of all chains blk.dat etc is closer to source, though a significant amount of data to review. Probably a few random samples would suffice.
The scripts which are used to run the tests being open source also helps, at least for those that understand it. I don't think we'll have issues with burden of proof if the right people with no vested interest confirm it's valididty. | 05:05:03 |
| komodobridge | [discord] <smk762> something like this, if it allows the blk.dat files, would be handy - http://yogh.io/#block:last | 05:06:38 |
| komodobridge | [discord] <smk762> something like this, if it allows importing blk.dat files, would be handy - http://yogh.io/#block:last | 05:06:52 |
| 16 Jul 2018 |
| komodobridge | [discord] <dwy> btw, maybe you can try this to test stuff : https://labs.ovh.com/kubernetes-k8s | 22:59:34 |
| komodobridge | [discord] <dwy> seems they provide alpha phase totally free | 22:59:50 |
| komodobridge | [discord] <dwy> don t know if that's stable, but for testing purposes maybe it's worth trying | 23:00:34 |
| 17 Jul 2018 |
| komodobridge | [discord] <polycryptoblog> 3rd party trusted verification you say, hmm may have the guys for that. | 01:48:10 |
| komodobridge | [discord] <blackjok3r> Yeah that would a good, if we can set up something like that. We are looking at an estimated 6-7TB of blockchain data generated in the test, but depending on cost, it could be much larger as I would prefer to make the test run longer that that if possible. | 01:52:23 |
| komodobridge | [discord] <blackjok3r> Yeah that would a good, if we can set up something like that. We are looking at an estimated 6-7TB of blockchain data generated in the test, but depending on cost, it could be much larger as I would prefer to make the test run longer than that if possible. | 01:52:36 |
| komodobridge | [discord] <blackjok3r> Btw, thanks @dwy I signed up for that OVH Alpha test, dobtful they have anything near the needed amount of CPU's but the way the test works, we can spread over loads of providers if need be. | 02:17:20 |
| komodobridge | [discord] <blackjok3r> Btw, thanks @dwy I signed up for that OVH Alpha test, doubtful they have anything near the needed amount of CPU's but the way the test works, we can spread over loads of providers if need be. | 02:17:31 |
| komodobridge | [discord] <Mylo(KomodoPioneers)>
https://cdn.discordapp.com/attachments/449949868904022036/468685021880778773/Metric_Definition_Proposal.pdf | 07:46:57 |
| komodobridge | [discord] <Mylo(KomodoPioneers)> This is a PDF print of a google doc shared by hyperledger.
Link from the working group home page (WG Working Documents) here https://wiki.hyperledger.org/groups/pswg/performance-and-scale-wg | 07:48:25 |
| komodobridge | [discord] <Mylo(KomodoPioneers)> From the last test @blackjok3r @webworker01 @smk762 - is it true to say:
```
- Live test to optimize smaller load
- 6k tx/s for 15 mins
- 4k tx/s for 30 mins
- 2k tx/s for 60 mins
- details: https://komodostats.com/scaling/index2.php?all
```
? | 08:04:44 |
| komodobridge | [discord] <Mylo(KomodoPioneers)> From the last test @blackjok3r @webworker01 @smk762 - is it true to say:
```
- Live test to optimize using a smaller load
- 6k tx/s for 15 mins
- 4k tx/s for 30 mins
- 2k tx/s for 60 mins
- details: https://komodostats.com/scaling/index2.php?all
```
? | 08:05:18 |
| komodobridge | [discord] <blackjok3r> It depends what you are trying to say... I can launch a test with 256 chains that cna run for 24H if thats a better meteric to use? | 08:06:06 |
| komodobridge | [discord] <blackjok3r> Maybe we should run some smaller scale tests for extended time periods, then a really large amount of chains for short time? | 08:06:37 |
| komodobridge | [discord] <blackjok3r> then we can prove that its technically possible to sustain very large numbers at scale... Its only a cost factor that limits the time. | 08:07:09 |
