From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: jehan@altheamesh.com Received: from krantz.zx2c4.com (localhost [127.0.0.1]) by krantz.zx2c4.com (ZX2C4 Mail Server) with ESMTP id 14aabd0f for ; Fri, 30 Dec 2016 01:15:51 +0000 (UTC) Received: from out1-smtp.messagingengine.com (out1-smtp.messagingengine.com [66.111.4.25]) by krantz.zx2c4.com (ZX2C4 Mail Server) with ESMTP id d229dfd8 for ; Fri, 30 Dec 2016 01:15:51 +0000 (UTC) Message-Id: <1483061047.444229.832649505.0712D346@webmail.messagingengine.com> From: Jehan Tremback To: jens , WireGuard mailing list MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Subject: Re: [fyi] test in the wild (wg + gretap + batman-adv) on tplink842 In-Reply-To: <1439626e-1f99-83fb-aab5-6a316346801e@viisauksena.de> References: <1482597351.888073.828789001.771C92C9@webmail.messagingengine.com> <1439626e-1f99-83fb-aab5-6a316346801e@viisauksena.de> Date: Thu, 29 Dec 2016 17:24:07 -0800 List-Id: Development discussion of WireGuard List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Got it, thanks. -- Jehan Tremback jehan@altheamesh.com On Sat, Dec 24, 2016, at 08:14 PM, jens wrote: > On 24.12.2016 17:35, Jehan Tremback wrote: > > Maybe I'm missing something obvious, but why would batman-adv affect the > > speed? > > not directly, > > or i don't understand where your conclusion came from. batman-adv means > a lot of little batman-mesh-protocoll packets (namely ogm packets, > protocoll traffic is around 300 kbits in our network. which also will > travel over gretap/wireguard) > > this scenario is equivalent to our free wifi network with 300++ nodes > and +-1000 users. > We use batman-adv as mesh protocol. While wireguard is layer 3 we need > to have layer2 over this tunnel. > so i tested wireguard + gretap + batman-adv successfully. > > compared to a other known layer2 crypted tunnel protocol (fastd) this is > 3 times faster, which is mainly due to wireguard is a kernel-module and > fastd lay in userspace. (-> context switche) > > i testet on weak devices, because this is what is deployed in big > numbers in our (and many other) Freifunk community. > also there are plenty of packets and stuff - so this explain the > relative low throughput. At least compared with clean linked test-setups. > > hope this explain my motivation and the reason why i test this in this > special scenario. > > Jens/fuzzle > > -- > make the world nicer, please use PGP encryption > >