3/20    Our current DSL service (bjt.net) sucks.  We need a good,
        high-bandwidth (1-2 megabits) DSL connection that has
        high uptime and a consistently low latency.  Anyone have
        good experiences?   -blojo
        \_ low latency's determined by the speed of light (I don't know
           of anyone who can change that yet) and distance to whatever
           you're trying to connect to.  So low latency to connect to
           what site?  And maybe queueing theory comes into play but that's
           a different story.  Latency usually isn't an issue in networks
           (except for UDP packets).  It's usually bandwidth that's
           important.
           \_ Hello, I am not a dumbass.  Our problem is that the ATM cloud
              we're fed into seems to be oversold.  So, yes, queueing theory
              comes into play here.  You need to have a higher level of
              expectation regarding the intelligence/experience of people
              on soda.  See oj's comment below.  -blojo
              \_ Say, blojo, do you have some sort of complex?  You blow up
                 everytime your overactive imagination perceives anyone even
                 slightly hinting that you may be a DUMBASS.
                \_ I don't have a complex insofar as I am aware.  I was
                   just astonished by the lameness of the post by someone
                   who seemed to think they just knew A WHOLE LOT.  Uhh
                   my imagination may be overactive but I tend to use that
          \_ Latency is HARDLY AT ALL determined by the speed of light. RTTs
                   energy imagining more interesting things than motd posts.
                      -blojo
                 \_ It's the twink-point effect, I'm telling you.
                    Just as intended.
          \_ Sign your posts, so everyone can know how dumb you are --oj
                \- when you start pushing ATM a lot of the abstraction
                breaks down and in specific areas the fundmaental technology
                difference matters. when you are pushing the switch, it may
                really start to matter how much bcast traffic or what kind
                of packet dropping heuristics they are using inside the
                switch to manage the buffering. this is hard even in the
                \_ Well, let's see.  If the speed of light were 200mph, then
                   I think you'd see a lot more latency.  (And find it very
                   hard to read, drive, fly, or many other things, but point
                   made via hyperbole).
                   \_ This is the real world we live in not your dream. In
                      reality the speed of light is not the limiting factor in
                      network speeds. That you felt the need to pull this 200
                      number out of the air merely shows how little your point
                      has to do with reality.
                packet switching world ... with the weird atm signalling
                protocols this stuff can be really hard to do right --psb
          \_ Latency is not determined solely by the speed of light.
          \_ Latency is NOT AT ALL determined by the speed of light. RTTs
             from here to the east coast are what, now, ~200ms?--but many
             times what it would take for light to travel there and back round
             trip. Latency is determined by the speed of switching
             electronics, which is much slower than light.
                \- in case your are interested, RTT to the east coast is
                higher than a lot of people think. say distance is 3000x2 mi
                and speed of light in glass is around 186288 x 60% =111800mps
                latency inside a switch is pretty small. most people use a
                send/recd buffer much smaller than the bw delay product so
                you are very much underfilling an empty pipe. --psb
                      \_ Actually, deep down the speed of light is the limiting
                         factor if you consider that the network switches
                         are literally a bunch of NMOS and PMOS transistors
                         hooked up together through polysilocon and metal
                         layers and that electron mobility is limited by
                         the speed of light which in turn limits the speed
                         by which logic circuits operate then yes, the speed
                         of light is a limiting factor.  Besides, if you read
                         carefully, the original poster never said that the
                         speed of light was THE SOLE limiting factor.
                         \_ "low latency's determined by the speed of light."
                            Looks like he came as close as you can to saying
                            that to me.
                            \_ Can you read or are you just and idiot? "low
                               latency's determined by the speed of light
                               AND however many hops you're trying to get
                               to."  I'm not going to quote exactly because
                               I'm too lazy to scroll up just for this stupid
                               post.
                                \_ Just an idiot.
                         \_ BINGO!  The other guy is a moron, though.  Of course
                            the speed of light has something to do with it.
                            What a serious super-duh to say it doesn't.  _That_
                            was the point I was making with the extreme 200mph
                            example.  It *matters*.  If sound went faster than
                            light, we'd be using that instead (somehow).  Join
                            the rest of us in the real world where the SoL does
                            matter.
                            \_ Did you pass high school physics?
                                \_ Sure did.  You pass HS English?
        \_ http://tycho.net serve your area?
        \_ FirstWorld (formerly slipnet) seems okay to me. BBNPlanet was pretty
           good too.