[Ndn-interest] Maximum Throughput in NDN
klaus at cs.arizona.edu
Mon May 20 12:02:56 PDT 2019
On 5/20/19 11:47 AM, Junxiao Shi wrote:
> Hi Klaus
> 1. What's the file size? Since there's less than 2 window decreases on
> average, I'd suggest using larger files to get more representative
> Each file is 100MB. There are six simultaneous flows, so total data
> amount is 600MB.
> Files were dd'ed from /dev/urandom.
> 2. What's the main conclusion? That getting data from CS is about 40%
> fast than getting it from the local producer app?
> I'm providing a data point, not drawing any conclusions.
> Also, I got a higher throughput (490 Mbps) on weaker hardware (i7-6600U
> laptop CPU) and not using the CS, with just 1 simultaneous consumer:
> i7-6600U has higher frequency (2.60GHz) than E5-2640 (2.50GHz).
But you have 12 CPU cores vs 2 on my laptop.
> Having only one flow does not reflect usual workload on a file server or
> a web browser client.
Yes, but 6 simultaneous flows should actually be *faster* in their
combined throughput. At least that's the case when link bandwidth is the
limiting factor. Moreover, you can distribute the flows among different
> Much lower RTTs too. Any explanation for the difference?
> I don't know. Everything is installed from APT without changes.
> Here are some more recent measurements from the last Hackathon, which
> achieve higher throughput than 25Mbps with similar RTTs:
> As you observed below, catchunks is much faster on the localhost vs.
> using a network link. I don't really know why that is, since in both
> cases (assuming link BW is high enough) CPU seems to be the limiting
> There could be packet loss using a network link, but no loss on Unix
> Yours, Junxiao
> Ndn-interest mailing list
> Ndn-interest at lists.cs.ucla.edu
More information about the Ndn-interest