Question about how Dash can scale

sha256sum

New Member
Jun 1, 2021
2
1
1
28
Hello there.

According to the docs, currently Dash is able to process around 56 transactions
per second (transactions per block divided by block time).

The blocksize is also dynamic, meaning it may increase or decrease according
to demand. How is the block size altered? If demand raises, how much time
will it take for the Dash network to adapt to it?

Even if the block size increase, would Dash network be able to handle thousands
of transactions per seconds? For example, 3000 transactions per second?

I really like Dash but I have some concernings about it being able to scale. We
can make instant payments using InstantSend, but I wonder if that would be
scalable if the network grew considerably.

I'll be thankfull if someone can answer me
 
  • Like
Reactions: Gass

sha256sum

New Member
Jun 1, 2021
2
1
1
28
Thank you GrandMasterDash, I'll watch the video and read the article soon as I have time
 

Geert

Member
Aug 26, 2015
259
82
88
IS can only process eighteen TPS, and this the "enhanced InstantSend" we're talking about. Until IS is upgraded or deprecated we can't scale.
 

GrandMasterDash

Grizzled Member
Masternode Owner/Operator
Jul 12, 2015
2,998
1,161
1,183
Please ignore @Geert he's feeling a bit pissy at the moment and having a tantrum with dash in general.

Anyone who has used other PoW coins knows dash is the easiest and fastest. Lightening is faster but comes with significant caveats. People give Lightening way too much credit and tolerance. Dash's Instant Send works, plain and simple. And despite what @Geert says, there is solid academic research and testing to scale when required. And if for whatever reason it does fail, you'd end up with a 2.6 minute block time. More so, because dash is based off bitcoin core, Lightening would also be possible.
 

Geert

Member
Aug 26, 2015
259
82
88
Pissy? I am having the time of my life! BTW, I was simply quoting what pasta said in his interview with TheDesertLoser.
 

Nthelight

Member
May 15, 2014
48
46
58
IS can only process eighteen TPS, and this the "enhanced InstantSend" we're talking about. Until IS is upgraded or deprecated we can't scale.
I raised the same concern in the Discord after seeing the interview with Pasta.
It's a legitimate concern and further optimizations are being looked at, but it seems the current design has caused some constraint somewhere.
Don't forget that the current design provides the ability to instantly respend, while the old didn't.
The current design is certainly an improvement in that regard.

Also, honestly, 17 TPS is still more than enough to support adequate growth for several years.
If we ever get to such high usage (well over 1 million tx per day), it will be an excellent problem for us to have.
 

AgnewPickens

Moderator
Moderator
Mar 11, 2017
435
162
113
57
I know some people would like to see DCG focus on L1 scaling more, Dash Platform as an L2 gives us a lot of data space, for things like games and profile data, and editing data for users, you don't want that bloating L1 as the verification layer. We can add a lot of value to the network on L2 without forcing scaling issues on L1.
 
Last edited:

xkcd

Active Member
Masternode Owner/Operator
Feb 19, 2017
253
225
103
australia
mnowatch.org
Dash Address
XpoZXRfr2iFxWhfRSAK3j1jww9xd4tJVez
DASH currently has 2MB blocks, they are not variable, but can be smaller if not full. The blocks come at rate of 2.625 minute intervals. A recent block I could find that was full had 5561 TXes in it, which gives a TPS of 35.

1623645575194.png

Now, the question of how many TXes the network can InstantSend is still an open one because we just released a new version v0.17 that has some performance improvements and we are yet to test it since we are waiting for the fork to hit, in a month or two. When it forks, we will test it on mainnet and confirm the numbers, but testnet shows the IS locking is able to keep up with block production.

In the case that the network is busy and you don't get a IS lock, then the TX remains unconfirmed in the mempool until it is mined, in exactly the same way as Bitcoin does. Meaning DASH's fallback is to be no worse than Bitcoin already is and just keep TXes in the mempool until a block is available. But unlike Bitcoin, DASH's blocks come 4x faster, so no matter how you look at it, DASH is faster.
 
  • Like
Reactions: Nthelight

Nthelight

Member
May 15, 2014
48
46
58
DASH currently has 2MB blocks, they are not variable, but can be smaller if not full. The blocks come at rate of 2.625 minute intervals. A recent block I could find that was full had 5561 TXes in it, which gives a TPS of 35.
Pasta did quote 17 TPS in his interview. That is low.

What does this mean?

What is the failure rate of IS locking?

Now, the question of how many TXes the network can InstantSend is still an open one because we just released a new version v0.17 that has some performance improvements and we are yet to test it since we are waiting for the fork to hit, in a month or two. When it forks, we will test it on mainnet and confirm the numbers, but testnet shows the IS locking is able to keep up with block production.
Can't wait to test the network. Hoping for positive results.

In the case that the network is busy and you don't get a IS lock, then the TX remains unconfirmed in the mempool until it is mined, in exactly the same way as Bitcoin does. Meaning DASH's fallback is to be no worse than Bitcoin already is and just keep TXes in the mempool until a block is available. But unlike Bitcoin, DASH's blocks come 4x faster, so no matter how you look at it, DASH is faster.
Sure, but this is not what we want, so we need to get a clear view on IS locking success rate at average usage and a full blown stress test :cool:
 
  • Like
Reactions: AgnewPickens