Blockchain Performance Update [1/3]

helium header

As members in the community have noticed, Proof of Coverage performance (and as an effect HNT rewards) has been notably different over the last few weeks. The core developers have been making lots of changes in order to try to keep the blockchain stable and producing blocks as the network scales. We wanted to give some insight on what we’re seeing, how this affects all blockchain participants, and what will be happening over the next few weeks.

Slower Block Times
As more Hotspots join the network, more poorly-performing Hotspots end up in the Consensus Group. Hotspots that are behind relays are much slower at networking, and can cause slower block times when elected to the Consensus Group. The result of slower block times is that it takes longer for HNT to be created and rewarded. The core team has made adjustments to timeout intervals and validation over the last few weeks in order to help reduce this problem and more stably produce blocks but we are also at the mercy of networking setups at 48K+ nodes.

Proof of Coverage Rate Interval
Last week we reduced the PoC challenge interval to 480 blocks from 240 blocks. At the target block time of 60s, this should mean that a Hotspot will be targeted every 8 hours (on average). However, due to slow block times, this interval could be as much as 12 hours or more. It won’t be uncommon to go a day or more without a single beacon, as PoC challenges can fail for a number of reasons. The rewards should be the same per beacon, but could take much longer to arrive. This will also mean you witness less often, as the Hotspots around you are also beaconing less frequently.


More Hotspots
Approximately 1000 Hotspots are joining the network each day. This reduces the per-Hotspot rewards, as the entire HNT pool is now split between more participants. It’s important to remember that this decrease is not linear. For example, adding 10% more Hotspots does not necessarily equal 10% less rewards, as it depends on how those Hotspots perform. If they have many witnesses, it could be significantly more than 10%. If they are poorly placed with no witnesses, they could have almost no impact at all.

Sync Issues
The community has also noticed a variety of sync issues over the past several weeks. The core developers have released several changes in order to address this including adding new download targets for snapshots (from Amazon S3), improving relays, and making some improvements to libp2p. We’ve also changed validation timeouts to support block production, as mentioned above, which should help all Hotspots. Since every Hotspot is a full blockchain node, they also have to validate all transactions when they absorb them into their local blockchain ledger. We believe that we’ve solved most of the systematic sync issues (only 1k hotspots across the 48k active seem to still be struggling at this point). Our best advice here is to ensure you are on a stable connection, get ahead of a relay by adjusting your network if you’re able, and if you continue to have sync issues, please contact your Hotspot manufacturer for support.

Coming Soon

The imminent move to Validators should help stabilize block times. Once we’ve seen Validators working well for a period of time we can look at increasing the PoC interval back to its old value of 240 blocks or less.

In the meantime, observing rewards and Hotspot behavior for longer periods of time helps individual Hotspot owners understand what’s going on in the system. Look at your Hotspot behavior over a week or more and ignore the 24 hour view. With the slower PoC rate 24 hours is no longer a useful sample size. The net amount of rewards per Hotspot should be roughly the same, but will now be spread out over a much longer period of time.

Thank you all for supporting the network as we scale!

Source :Discord official channel

Related Post