@paniq @cmuratori @Jonathan_Blow @nothings If you don't have congestion avoidance, you will fail at scale. "Just send packets" doesn't work.
-
-
Replying to @jwatte
@jwatte@paniq@Jonathan_Blow@nothings Congestion avoidance works just as well with connectionlessness (http://curvecp.org/decongestion.html …)1 reply 0 retweets 2 likes -
Replying to @cmuratori
@cmuratori@jwatte@paniq@Jonathan_Blow 0% of that article is about connectionless. Scheduling only makes sense with connections.2 replies 0 retweets 0 likes -
Replying to @nothings
@nothings@jwatte@paniq@Jonathan_Blow It is _entirely_ about connectionless. It's a discussion of CurveCP's congestion handling.2 replies 0 retweets 1 like -
Replying to @cmuratori
@nothings@jwatte@paniq@Jonathan_Blow CurveCP is entirely UDP-based.1 reply 0 retweets 1 like -
Replying to @cmuratori
@cmuratori@jwatte@paniq@Jonathan_Blow CurveCP is a protocol for creating secure connections over UDP.1 reply 0 retweets 0 likes -
Replying to @nothings
@cmuratori@jwatte@paniq@Jonathan_Blow If I implement an exact clone of TCP/IP on UDP, that doesn't make my protocl connectionless.2 replies 0 retweets 1 like -
Replying to @nothings
@nothings@jwatte@paniq@Jonathan_Blow Not sure how that is relevant? The point is just that you do not need TCP for congestion control.2 replies 0 retweets 2 likes -
Replying to @cmuratori
@nothings@jwatte@paniq@Jonathan_Blow Nor do you need guaranteed delivery to have congestion control. Obviously there is always a...1 reply 0 retweets 1 like -
Replying to @cmuratori
@nothings@jwatte@paniq@Jonathan_Blow ... "connection" of some kind if you're going to send more than one packet or expect acks!2 replies 0 retweets 1 like
@nothings @jwatte @paniq @Jonathan_Blow There is no such thing as "connectionless" if you want to take the term literally.
-
-
Replying to @cmuratori
@cmuratori@jwatte@paniq@Jonathan_Blow Nobody claimed you need TCP for congestion control. Claim was that NO congestion control = bad.1 reply 0 retweets 0 likes -
Replying to @nothings
@nothings@jwatte@paniq@Jonathan_Blow No. The argument originally was about whether TCP was well-designed, and we are still on that.1 reply 0 retweets 1 like - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.