Yesterday we launched our latest testnet to analyse the effect of encryption on node memory performance. This builds on the previous QuicNet, which saw a move from TCP to QUIC, and drew a rapturous response.
QUIC is evidently the future. Nevertheless, this being cutting-edge engineering, for every step forward the law says there has to be half a step back ...
Welcome to the last update of the year. We’re happy to be finishing on a high, with the CloserNet testnet and the community helping us check our fix for what we hope will prove to be the last major bug in basic data handling. We were making an assumption that nodes were being ordered by closeness to our node, whereas this was...
This week we’ve all hands on deck, filling holes, bashing in nails and replacing rotten timbers in the previous testnet so we can launch the thing of beauty that is the ReduceConnectionsNet, which thus far seems to be sailing along pretty nicely.
This week we’ve been further considering the economic on and off ramps to the Network—particularly in the early stages after genesis.
We’ve also been looking at the underlying cause of excessive open connections. Having stopped FloodSub the current testnet is better than the last in that regard, but heaptracking still reveals far more connections than we need.
The most important news this week is the winding down of the Bittrex exchange. Anyone who has MAID on #Bittrex needs to take it off the exchange sooner rather than later.
Other news is that the latest testnet seems to have solved some problems and unearthed others in a one-step-forward-one-step-back kinda style.
It prompted me to make a topic for #socket on the #SafeNetwork forum and #MaidSafe's founder has made a few comments there so you might want to take a look:
Well, it was bound to happen some time. After a run of successes, the RoyaltiesPaymentNet was cursed by high mem which killed off many nodes and left the rest pretty much zombified.
Spookily all was fine on our internal testnets (albeit with some slightly raised memory levels). Could poor RoyaltiesPaymentNet have been struck down by dark forces beyond our ken? 👻
Well, perhaps there's a scientific explanation... Read on
We’ve often sympathised with Sisyphus in the Greek legend. Sisyphus had to push a boulder up a hill for eternity only for it to go rolling back down again as soon as he neared the summit.
Not wishing to tempt the notoriously vengeful Greek gods, but we’re increasingly certain that this time we’ve finally cracked it, and we’re sitting pretty on the plateau. Why the confidence?
We’ve had two testnets on the go in the last week, aiming to clarify issues in replication and at the client, as well as investigating the mem increases we’ve seen across nodes.
So far, while we’re still seeing client issues, we’ve been able to clarify what’s going on there and so fixes to failed CashNote reads, and other upload issues, are in the works.
We’re looking more concretely at the client. The last testnet aimed to see how clients performed with the concurrency and batch-size arguments, but a bug in StoreCost retrieval scuppered that effort.
Here we want to:
Confirm our fix is in place for that.
Explore how concurrency and batch size effect uploads.
Look at how Kademlia caching effects downloads of popular data.
We’re making excellent progress, and a back-of-a-beermat calculation suggests we’re past the halfway point to Beta now, with payments, store cost and rewards all in place.
Call it synchronicity, a confluence of decentralised thinking, or longstanding issues coming to a head, but just as @dirvine was chatting to the team this week about protection against Sybil attacks, a post popped up alerting us to Vitalik Buterin’s thoughts on the topic.
David’s initial response, is ion teh forum. Tldr; outside of blockchain-land Sybil may not be so scary after all.
After a successful DialNet we’re looking to test our reworked payment process. Now instead of badly guesstimating costs on the network, we ask each and every node how much they’d like and pay that. (This can be tweaked to avoid bad actors easily enough). As such, every node will receive rewards for PUTs directly!
Safe Network Update 14 December, 2023 (safenetforum.org)
Welcome to the last update of the year. We’re happy to be finishing on a high, with the CloserNet testnet and the community helping us check our fix for what we hope will prove to be the last major bug in basic data handling. We were making an assumption that nodes were being ordered by closeness to our node, whereas this was...