From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jun Xiao Subject: Re: [ovs-discuss] vswitches performance comparison Date: Wed, 22 Jul 2015 02:48:51 +0800 Message-ID: <9D0E6ED2-6171-4EF5-AD21-01B1844B5136@cloudnetengine.com> References: <----Tc------lRRzc$3e501353-5ebe-4161-b9d4-01ebdf81a6de@cloudnetengine.com> <738D45BC1F695740A983F43CFE1B7EA92E2BEF4B@IRSMSX108.ger.corp.intel.com> <738D45BC1F695740A983F43CFE1B7EA92E2BEFD4@IRSMSX108.ger.corp.intel.com> Mime-Version: 1.0 (1.0) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable Cc: dev , discuss To: "Gray, Mark D" Return-path: Received: from out1134-201.mail.aliyun.com (out1134-201.mail.aliyun.com [42.120.134.201]) by dpdk.org (Postfix) with ESMTP id 064615A43 for ; Tue, 21 Jul 2015 20:48:53 +0200 (CEST) In-Reply-To: <738D45BC1F695740A983F43CFE1B7EA92E2BEFD4@IRSMSX108.ger.corp.intel.com> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" > On Jul 22, 2015, at 2:36 AM, Gray, Mark D wrote: >=20 >=20 >>=20 >> I'd like to hope that's my methodology problem, but I just follow the >> installation guide without any customization. >>=20 >> Hi Mark, do you have any performance data share with us? Maybe we are >> using different type of workloads, like I mentioned I am using typical da= ta >> center workload, I guess you are talking about NFV type of workload? >=20 > The number getting floated around on the mailing list recently is 16.5Mpps= > for phy-phy. However, I don't think we have any iperf data off-hand for yo= ur > usecase. When we test throughput into the vm we usually generate the traff= ic externally > and send NIC->OVS->VM->OVS->NIC. This is a little different to your setup.= >=20 I guess pmd driver is used inside VM in that case, right? > I do know, however, that ovs-dpdk typically has a much larger throughput t= han > the kernel space datapath. >=20 I'd like to say it depends on workloads, for small/medium packet size worklo= ad, that's definitely true, while for TSO size workload, it's not that obvio= us (or worse) as data path overheads are amortized and H/W can be leveraged.= > Have you seen this? https://wiki.opnfv.org/characterize_vswitch_performanc= e_for_telco_nfv_use_cases >=20 Thanks for the pointer, I'll try later. >>=20 >> Thanks, >> Jun