Hi everyone 👋
As part of the shift from Consul we have had to rebuild the service management layer in Stargate in order to handle service metadata, something required to determine which services are hosting apps in order to define DNS routing correctly across the network.
This means that on top of the KV changes discussed in earlier updates, we now have:
- Service registration (RPC)
- Service deregistration (RPC)
- Service health (RPC)
- Service metadata (RPC)
This is a significant move in the entire consul migration work stream. It should be ready to move out of test.network to the mainnet at the end of this month.
As part of this work we’ve also built in more intelligence to the network services layer, for example Stargates reporting on the Gateways connected to them directly:
Arthur spoke about the why of moving on from Consul in yesterday’s AMA:
“Network coordination and configuration sharing was and still is one of the key components to Edge. At the very beginning there was an immediate requirement to implement a solution that would last us long enough to work out exactly what we needed from it.
“We introduced Consul early (not long after ditching ZooKeeper, but that’s a story for another time). As development progressed, it became clearer that what we needed from a decentralised network manifest was not always going to be covered by Consul, but introducing it was a level of technical debt we could, on balance, afford.
“Predictably, it became sluggish when trying to handle the increased number of healthchecks as more devices connected. Unpredicabtably, it also didn’t cope with regular KV writes, so as more users onboarded with CDN we saw a rapid slowdown of deployment. When I say rapid, I mean it took something like 30 seconds to deploy a new instance of CDN, which might seem pretty fast for global deployment, but we could see it wouldn’t scale.
“Over the last 12 months we’ve raised a number of issues relating to performance without adequate feedback. We’ve also seen a rather quick shift to monetisation methods that we don’t necessarily agree with, including the removal of features for non-enterprise users. To be an enterprise user we would need to host via Hashicorp owned infrastructure, which means it is no longer decentralised within our network.
“_For us, with all of the existing performance issues it is not yet a valuable option, so we took this opportunity to replace it with a system we’d been planning for far later in the roadmap._“
We’ve finished the last stages of the watermark feature for CDN, which has now been released to Testnet. The next major feature is fallback image support which will enable the return an image instead of an error message when a source file can’t be found or processed.
We’re involved in a pitch for a infrastructure project for a Gov. department here in the UK. It’s very early in the process, but there is great alignment with our technology.
We had a two hour AMA with Jason from TNC yesterday. Amongst other things he answered questions on future mergers, tokenonmics and the various businesses that make up the TNC Group. You can see a full transcript here:
In summary: we’re on for 01st October for unlock.
We kicked off a partnership with a very promising blockchain company, specifically exploring how we can leverage their technology in the network to reduce fees and increase payout times. Separately to this we’re moving their site and explorer in to the Edge network. As soon as this is up - which should be this side of Christmas - we’ll share it with the community.
On new business efforts in general, Chris had this to say in the AMA yesterday:
“_It’s early days and we have a bunch to complete in support (the site update for e.g.), but the strategy is in place and is being executed on. We’re starting to see green shoots and there will be updates about new engagements as we move forward. We’re obviously focused on CDN in the main, but Edit is also a core focus for me. It’s a powerful set of tech and does a great job of demonstrating what the network is capable of._“
The market opportunity for CDN is highly significant. It was valued at 12 billion US in 2019, with well over a million paying customers in the space, and it is expected to reach a value of 50 billion by 2025.
After many years of development production testing it’s great to be in the midst of taking the tech to market. There’s so much scope for the platform and we’re really only just starting to scratch the surface.
Anyway, that’s it for this week!
Enjoy your weekends 👌