Latency Benchmarks for the Network

Continuing the discussion from MaidSafe Dev Update 8th December 2015:

Hey @Ross, @Fraser

I know that this weeks news was a little disappointing.

However, the week prior you guys indicated you were running some tests.

Were you able to pull any latency benchmarks from those tests? Will such benchmarks be available prior to the public test network (notice I am not asking when that will occur)?

The latency benchmarks are hugely important to all kinds of applications, in particular gaming and Real-Time Collaborative Editing environments, so I hope that the team posts that sort of information as soon as available.

On that note, is the routing portion the key factor in latency? How much does Crust or other pieces of the network play a role?

5 Likes

Really hopeful for this!

I wanna see a SAFEtube and lots of SAFE multiplayer games :smiley:

1 Like

I don’t think that is possible with safe. Latency will be very high due to hops.

That will not be a problem due to SAFEs parallel access of the data. All players do buffering so multiple chunks would be requested when the player starts and when the 1st chunk arrives the others will be on average be there too.

There is other threads that discuss the issues with multiplayer games. I guess for the fast action ones, we will see a hybrid solution involving direct connection between players and/or the current gaming system. Hops is the death of extremely low latency needed for some games, so hybrid is the way to go.

4 Likes

I have just pinched this link from another thread, but applies well here too, I think.

2 Likes

Do we have an update on this? Would a 5G context where latency may be sub ms make a difference in the hop or mesh context in aggregate latency in most use cases?