True! On Dec 6, 2015 6:11 PM, "Josh Luthman" <j...@imaginenetworksllc.com> wrote:
> Correct. Let's be honest, it's all TCP Netflix... So better :) > > Josh Luthman > Office: 937-552-2340 > Direct: 937-552-2343 > 1100 Wayne St > Suite 1337 > Troy, OH 45373 > On Dec 6, 2015 7:09 PM, "Mathew Howard" <mhoward...@gmail.com> wrote: > >> So actual throughput could be slightly better or slightly worse, >> depending on the type of traffic? >> On Dec 6, 2015 5:00 PM, "Dan Sullivan" < >> daniel.sulli...@cambiumnetworks.com> wrote: >> >>> Hi, >>> >>> >>> >>> Josh is right. >>> >>> >>> >>> 2.5 msec frame also was created to provide much lower latency for fixed >>> ratios: 75/25, 50/50, 30/70. >>> >>> >>> >>> The headline throughput will be a little less, because there is more >>> frame overhead since double the amount of packets / frames, but overall >>> latency is decreased which can help TCP and also increase TCP throughput. >>> Throughput for 2.5 versus 5 msec should be compared for each case. >>> >>> >>> >>> Dan Sullivan >>> >>> ePMP Software Manager >>> >>> >>> >>> *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Josh Luthman >>> *Sent:* Saturday, December 05, 2015 12:57 PM >>> *To:* af@afmug.com >>> *Subject:* Re: [AFMUG] ePMP 2.5ms frame >>> >>> >>> >>> Absolutely lower latency. Should be identical to pmp100 I think? >>> Definitely a lot better than the 5ms. >>> >>> >>> >>> >>> Josh Luthman >>> Office: 937-552-2340 >>> Direct: 937-552-2343 >>> 1100 Wayne St >>> Suite 1337 >>> Troy, OH 45373 >>> >>> >>> >>> On Sat, Dec 5, 2015 at 1:55 PM, Adam Moffett <dmmoff...@gmail.com> >>> wrote: >>> >>> Yeah, but I haven't had to colo it with a PMP100 so at the time I >>> glossed over it. >>> >>> I was wondering if I was missing out on lower latency.....sounds like >>> that's a yes. >>> >>> On 12/5/2015 10:22 AM, Gino Villarini wrote: >>> >>> wasnt 2.5 frames announced some time ago? >>> >>> >>> >>> On Fri, Dec 4, 2015 at 9:15 PM, Josh Luthman < >>> j...@imaginenetworksllc.com> wrote: >>> >>> Like 80% sure I'm right but ya latency is the most important. >>> >>> Josh Luthman >>> Office: 937-552-2340 >>> Direct: 937-552-2343 >>> 1100 Wayne St >>> Suite 1337 >>> Troy, OH 45373 >>> >>> On Dec 4, 2015 7:47 PM, "Mathew Howard" <mhoward...@gmail.com> wrote: >>> >>> Isn't throughput supposed to be slightly worse with 2.5ms, or am I >>> remembering that wrong? There is definitely a big improvement in latency. >>> >>> Either way, I don't think the difference in throughput is enough to care >>> about, I'd rather have the lower latency. >>> >>> >>> >>> On Fri, Dec 4, 2015 at 1:59 PM, Josh Luthman < >>> j...@imaginenetworksllc.com> wrote: >>> >>> A bit more throughput. Not worth changing from 5ms to 2.5ms if all >>> you're after is throughput. Any new APs I do 2.5ms instead of 5ms since >>> there's no reason to do it "the old way". You do get better latency - >>> 7.5ms improvement. >>> >>> >>> >>> >>> Josh Luthman >>> Office: 937-552-2340 >>> Direct: 937-552-2343 >>> 1100 Wayne St >>> Suite 1337 >>> Troy, OH 45373 >>> >>> >>> >>> On Fri, Dec 4, 2015 at 2:56 PM, Adam Moffett <dmmoff...@gmail.com> >>> wrote: >>> >>> It's a big day for ePMP threads. >>> >>> If I recall correctly, the 2.5ms frame size was introduced to aid in >>> collocation with the PMP100. Is there any performance impact? My first >>> thought was that a shorter frame would give me lower latency, but it's >>> probably not that simple. >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>