[ANN][SIM] Simcoin - A Simple Coin

[NxtChg]

Progress report

As you know, this week I was working on the transaction downloader. Here's a new version of it, which is almost a complete rewrite:

https://simcoin.info/bootstrap-test2.zip

It's now more efficient, better tolerates lost packets and the overhead is dramatically reduced.

It also downloads the file twice, in two separate streams, to demonstrate scalability. Each stream downloads at the same speed as the single stream version, so the total speed is doubled.

--

As usual, most of the time was spent trying to foresee any possible issues or attacks. This is the most daunting part of the project, but since so few developers seem to do it, I think I must continue doing so.

[grandpa_seth]

This should be the Simcoin logo.

[NxtChg]

Progress report

This week I continued working on bootstrapping/downloading, because before you can start downloading, you need to figure out what you need to download and get a consensus on that.

So there are 2 more tasks on top of the downloader. Can't go into details here, but the previous solution required an additional database table and the new one works with existing structure, and in general is more flexible.

Also did more SQLite performance tests to make sure the new scheme is fast enough.

In a day or two I expect to finish this part of the system and hopefully never touch it again

This time the solution seems to be quite good.

[mcjavar]

Progress report

This week I continued working on bootstrapping/downloading, because before you can start downloading, you need to figure out what you need to download and get a consensus on that.

So there are 2 more tasks on top of the downloader. Can't go into details here, but the previous solution required an additional database table and the new one works with existing structure, and in general is more flexible.

Also did more SQLite performance tests to make sure the new scheme is fast enough.

In a day or two I expect to finish this part of the system and hopefully never touch it again

This time the solution seems to be quite good.

Do you have some stats about the distribution?

[NxtChg]

Do you have some stats about the distribution?

That's a good question! I haven't done this for a while. So here goes:

Distribution update



234 accounts own simcoins now vs 210 after the IPO and vs 254 in August.

So we lost 20 people, who exited completely. That's a bit insulting. Couldn't you leave at least 10 K, in case I turn out not to be the complete failure you think I am?

But, oh well, hopefully the joke will be on them.

The top 10 accounts own 23.42% vs 21.26% on August 31.

You can see it in the graph, the money went from the bottom half of the curve to the top half.

Rich get richer, you know

While the bottom half complains that "we're 99%, give us more", as if they have anyone else to blame but themselves...

[rlh]

Ha! Funny image progression, but your last image is wrong.  Should have been this:

[NxtChg]

I thought I also give you an update on what I am working on right now.

A big area, which is not yet finished, is reorg. Although the consensus mechanism should be very reliable, there might still be situations, particularly in catastrophic scenarios, where some nodes will have a different view of transaction history and need to reorganize it.

The Reorg Boundary is a point in time beyond which no reorg is possible. Nodes will insist on their view of the history no matter what. It should be some reasonable value and I think 4 hours is a good point.

Now, the lazy solution to deal with reorg would be to just flush everything from the point of fracture and re-download it again. The problem, though, is that at 100 TPS 4 hours can amount to about 200 Mb and at 1,000 TPS it can be 2 Gb!

Granted, this should be extremely rare, and we will probably be at 10 TPS, not 100, for quite some time, so it's more like 20 Mb of transactions, tops.

So here's my current dilemma – be lazy or try to invent some complicated scheme to figure out what changed and then rollback/re-download/apply the differences.

It's quite stupid to re-download everything just because one transaction might be different, but detecting what exactly is different over a hostile and unreliable network, and then merging the changes into main "branch" of history, is also far from trivial, actually bordering on "insanely complicated"

Still thinking about this...

[wizzardTim]

Please don't take the lazy approach. I mean since you dedicated so much work and currently SIM seems to me that is the only cryptocurrency that we can rely for building something totally new, innovative and stable. Every minute dedicated to research before implementation is of utmost importance, since we avoid future impediments.  

[NxtChg]

Please don't take the lazy approach. I mean since you dedicated so much work and currently SIM seems to me that is the only cryptocurrency that we can rely for building something totally new, innovative and stable. Every minute dedicated to research before implementation is of utmost importance, since we avoid future impediments.  

Thanks, but this project is behind schedule and a lot of people want to see some tangible results already

So I have to cut a few corners to try and deliver sooner. Mostly things that can be improved later, when we need to handle higher volume.

For reorg I will probably use a combination of two schemes: flush everything, but try to re-download only the relevant section of the history and thus cut the download size about 10 times. Half-assed approach

Besides, this should never happen anyway, it's only for very rare cases.

[cryptodevil]

FWIW, I'm happy to wait.

This is one of the few projects currently in crypto that I don't doubt is going to come through in the long run and I would rather you don't feel pressured to rush out solutions to problems.

[NxtChg]

FWIW, I'm happy to wait.

This is one of the few projects currently in crypto that I don't doubt is going to come through in the long run and I would rather you don't feel pressured to rush out solutions to problems.

Thanks, guys

[lyynx]

FWIW, I'm happy to wait.

This is one of the few projects currently in crypto that I don't doubt is going to come through in the long run and I would rather you don't feel pressured to rush out solutions to problems.

Thanks, guys

Same here, I am waiting patiently, quality comes first. Keep up the great work.

[Nxtblg]

FWIW, I'm happy to wait.

This is one of the few projects currently in crypto that I don't doubt is going to come through in the long run and I would rather you don't feel pressured to rush out solutions to problems.

Thanks, guys

Same here, I am waiting patiently, quality comes first. Keep up the great work.

Me three. 

[KLONE::Vader]

your minions are well behaved 

[NxtChg]

your minions are well behaved 

For that, I will reward them generously.

[KLONE::Vader]

cross fingers

[grandpa_seth]

your minions are well behaved 

For that, I will reward them generously.

And what does a good little minion or Wayne's World fan say to that? We're not worthy! We're not worthy!

[NxtChg]

Progress report

Slow week, had to take care of some real life stuff.

So, as you know, I finalized bootstrapper this week, plus reorg. The whole architecture starts to look very simple and elegant.

There are 3 major parts: Transaction Gobbler, Bootstrapper (which now also handles reorg) and the Heart, which handles transaction consensus.

All of them are practically done, yet my to-do list still has a gazillion of small tasks This is the proverbial last 10%, which take another 90% of the time Anyway, we will get there one day.

Also rewrote this piece of garbage: https://github.com/ghedipunk/PHP-Websockets and now I have a nice, simple, non-blocking websockets server in PHP, which can be used for various dynamic web pages: exchange, explorer, web wallet, you name it. Even games

[gs02xzz]

Stellar revealed a flaw in Ripple/Stellar consensus. It seems there are still lots of room for better consensus.

https://www.stellar.org/blog/safety_liveness_and_fault_tolerance_consensus_choice/

Prof. David Mazières, head of Stanford’s Secure Computing Group, reviewed the Ripple/Stellar consensus system and reached the conclusion that the existing algorithm was unlikely to be safe under all circumstances. Based these findings, we decided to create a new consensus system with provable correctness. This effort, led by Prof. Mazières, is underway. His white paper and the accompanying code are expected to be released in a few months.

[NxtChg]

Stellar revealed a flaw in Ripple/Stellar consensus. It seems there are still lots of room for better consensus.

https://www.stellar.org/blog/safety_liveness_and_fault_tolerance_consensus_choice/

Wow, another one bites the dust, and a big one!

Well, can't say this is a big shock, because when I've read their full paper I was surprised at how bad their system actually was - it made so many assumptions about the state of the network in order to function correctly.

We are still investigating the triggers for this consensus failure, but believe it is caused by the innate weaknesses of the Ripple/Stellar consensus system outlined above compounded by the number of accounts in the network. Presently, we have approximately 140,000 active accounts a week and over 3 million total accounts which is in excess of the approximately 120,000 total accounts (active and inactive) this stack has previously supported.

Our monitoring of the network has made it clear that the underlying Ripple/Stellar consensus system is not performing at this level of scale, which is still small relative to the global financial system. In order for such protocols to perform at real-world levels with the expected degree of safety, this number of accounts should not be a problem.

Just wow.

This also shows how difficult the distributed consensus is. Although I hope to do better, every such failure makes me suspect that I am deluded (Sell, sell, sell! )

Also, I can't fall back to Ripple's consensus now if Simcoin's consensus fails, at least not until they invent something better.

Well, let's hope it will work.