From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Marek Lindner Date: Sun, 16 Nov 2014 14:08:43 +0800 Message-ID: <3113299.290J3YiqPU@diderot> In-Reply-To: References: MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart1552012.rIisWV2ht0"; micalg="pgp-sha1"; protocol="application/pgp-signature" Subject: Re: [B.A.T.M.A.N.] wired only 10 gigabit batman-adv mesh Reply-To: The list for a Better Approach To Mobile Ad-hoc Networking List-Id: The list for a Better Approach To Mobile Ad-hoc Networking List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Mehul Sharma , 'The list for a Better Approach To Mobile Ad-hoc Networking' --nextPart1552012.rIisWV2ht0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="us-ascii" Hi, > Good Afternoon from Boston. I really love Batman-Adv= ... > brilliant layer 2 functionality. >=20 > I want to use batman-adv in a wired (gigabit and 10-gigabit) only mes= h and > wanted to know your insights. makes me happy to hear you love our project. Typically, we communicate = via our=20 public mailing list allowing various sources to chime in at any point. = Since I=20 don't see any reason for privacy I am cc'ing the mailing list in my ans= wer. > The example case scenario is as follows: >=20 > 1) 4 to 6 AMD servers with 6 10-Gigabit NICs each. >=20 > 2) 2 or 3 10-Gigabit NICs used for batman-adv, which are then con= nected > in ring or torus topology directly (no external switch involved) >=20 > 3) the remaining interfaces on the server are connected to the LA= N > (switches, routers etc) >=20 > 4) the virtual machine (qemu-kvm) tap interfaces, the physical > non-batman-adv ethernet and bat0 interfaces are put in a bridge (brct= l), so > now we have the ability for virtual machines, wired hosts on the lan = to go > via batman-adv and talk to each other. >=20 > Is there any, down size to doing this? I see at the most 2 - 100 serv= ers in > one network.... >=20 > From what i understand: >=20 > 1) that the live migration of virtual machines (qemu-kvm) will be se= en > just as a migrating non-mesh client so my assumption is that live mig= ration > should work from that perspective. Also, what if the tap interfaces o= f the > virtual machines are given to bat0 itself (if it might help in live > migration / increasing throughput) ? >=20 > 2) The MTU if set for 1500 or 9000 or higher (eg barman-adv reads -- > "define ETHERMTU ETH_DATA_LEN") would be taken automatically by batma= n-adv > and anything below 1500 would be fragmented, which gives me the idea = that > higher MTUs would not be a problem for batman-adv to handle. >=20 > 3) There is no restriction to the number of clients in batman-adv. >=20 > am i somewhat close in understanding batman-adv? .... apologies if no= t... >=20 > Also would layer 2 forwarding by batman-adv would be close, same or b= etter > when compared to bridge (linux brctl) packet forwarding? >=20 > I have built converged-unified distributed qemu-kvm system (all metad= ata > less design, with web-interface and cli, quite the opposite of vmware= and > open-stack type centralized approaches) and was in the preliminary st= age of > looking at the possibility of integrating batman-adv into the design.= >=20 > Your input will be valuable for me to give server and desktop > virtualization a mesh architecture on top of already distributed desi= gn. I keep your description intact to allow other people to comment as well= . Before we dive into the batman-adv details I'd like to understand what=20= advantage batman-adv brings to the table in your scenario. The batman-a= dv=20 project aims to facilitate layer2 routing in primarily wireless setups = with=20 dynamically changing links due to link quality changes or links being m= odified=20 in an uncontrolled fashion (community mesh network). While batman-adv a= lso is=20 able to run on wired backbones this never was the main target and bears= a=20 number of drawbacks compared to other technologies. A simple example to= =20 picture this: The standard Linux bridge (configurable via brctl) does n= ot run=20 any link layer protocol to estimate the quality of one link compared to= =20 another. This will give you huge advantages in terms of overhead with t= he cost=20 of all links being treated equal. While this work fine on an all-wired = setup=20 it represents an unacceptable trade-off for wireless networks. From=20what I can gather you are not running wireless but high throughput= wired=20 links. What has brought you to batman-adv ? Cheers, Marek --nextPart1552012.rIisWV2ht0 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: This is a digitally signed message part. Content-Transfer-Encoding: 7Bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAABAgAGBQJUaD9uAAoJEFNVTo/uthzA2+sH/2/V5mdy5SLrmP6xf2V878MV F7UZueQQCY7GuwILpCVSX43L2bhVHXsLdIuagPMRgb33Ox5cFBXGzhgSbFppUcgJ 2Ct6bgbrQOT53vwfYSuRAv/uNxWerFehShYiv/A75TMnmKZ/52fo4DZd6Q14D57G qWeSQxtEOqb31I82M1kNHxkDFxMG/qjtsQw9WQdwNTioM0u+nddVorqqjF3VrkSX KG9h7TUDpRsoRf3vSPkRggVxQ1AR9i+Kx0FHdwXXvVZZP6oOdVYhFFCfI/BpQu4F KsOTZ80Dx1wrJ4ihVlKGyuQltGnLGsd01dTz31UsNB4QNT5sphLQBpafLc2+pv4= =CY5a -----END PGP SIGNATURE----- --nextPart1552012.rIisWV2ht0--