They're completely different CPUs, so even given the same number of cores, 1.4ghz isn't necessarily faster than 1ghz (I have no idea how the two actually compare, but I suspect the CCR is faster).
One thing to keep in mind with the 3011 vs CCR, is that on a CCR every port has a full 1gbps connection to the CPU, but on the 3011, the ports are divided into two switch groups, and each switch group just has a single 1Gbps connection to each CPU, which is shared between those five ports (so it can theoretically router 2Gbps per switch group, but that also changes a bit if the SFP port is in use). That makes the maximum your can route through a 3011 about 4Gbps (and that assumes that you have everything connected to all the right ports), but even the smallest CCR is capable of routing about 8Gbps (across 8 1gig ports). On Tue, Jul 24, 2018 at 10:25 AM, Kurt Fankhauser <lists.wavel...@gmail.com> wrote: > If a tower router is only doing OSPF routing and no connection tracking is > a 3011 faster for that 1 task than a CCR since it has 1.4Ghz CPU vs the CCR > with a 1Ghz CPU? > > On Tue, Jul 24, 2018 at 11:22 AM, Kurt Fankhauser < > lists.wavel...@gmail.com> wrote: > >> If a tower router is only doing OSPF routing and no connection tracking >> is a 3011 faster for that 1 task than a CCR since it has 1.4Ghz CPU vs the >> CCR with a 1Ghz CPU? >> >> On Fri, Apr 6, 2018 at 11:42 AM, Adam Moffett <dmmoff...@gmail.com> >> wrote: >> >>> My rule of thumb is >50% = upgrade. >>> I wouldn't argue with 25% either. >>> >>> We see a % CPU usage which updates every second. I don't know if it's a >>> 1 second average, or a rolling 5 second average, or whatever. Either way >>> there could be milliseconds within that average where all CPU time is >>> occupied and it wouldn't show in the display. If we're trying to carry >>> traffic which is sensitive to milliseconds of delay (VoIP) then my feeling >>> is ought to be safer to keep the CPU usage low. >>> >>> I have no data to back this up. It's basically just the philosophy that >>> I'd rather have it and not need it than need it not have it. >>> >>> >>> >>> ------ Original Message ------ >>> From: "Mathew Howard" <mhoward...@gmail.com> >>> To: "af" <a...@afmug.com> >>> Sent: 4/5/2018 9:10:41 PM >>> Subject: Re: [AFMUG] CPU - when to upgrade >>> >>> I start getting paranoid if I see more than about 25% CPU load on a >>> Mikrotik... whether or not that's actually having any impact on >>> performance, I don't know, but I suspect at 70% it is - I would certainly >>> upgrade it to something faster at that point. What model of Mikrotik is it? >>> >>> On Thu, Apr 5, 2018, 7:24 PM Rhys Cuff (Latrobe I.T) < >>> r...@latrobeit.com.au> wrote: >>> >>>> Hi Guys >>>> >>>> >>>> >>>> When you have a router and cpu is hitting about 60 – 70% with traffic >>>> load would that impact speeds? >>>> >>>> I’ve got a few links that don’t seem to go as hard as they should, but >>>> I can’t find the reason. >>>> >>>> The routers are MikroTik and the radios (also highish cpu) are ubnt >>>> with plenty of airtime free) >>>> >>>> Hmmm, That’s another question, at what Airtime % will you start to see >>>> impacted performance? >>>> >>>> >>>> >>>> >>>> >>>> Thanks >>>> >>>> >>>> >>>> Rhys >>>> >>>> >>>> >>> >> > > -- > AF mailing list > AF@af.afmug.com > http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >
-- AF mailing list AF@af.afmug.com http://af.afmug.com/mailman/listinfo/af_af.afmug.com
