<p>der 12.2 ist übrigens am donnerstag*g<br>und ist wiedermal furchtbar kurzfristig weil wir so furchtbar motiviert sind,.. *g<br></p><p>und wer nun überhaupt nicht mehr versteht, um was es hier geht, und was der freebone ist,..<br>
</p><p>keine Sorge das wird bald genauer dargelegt, nur hab ich grad nicht die Zeit das zu tun,..</p><p>lg Markus</p><p>und sorry das untrige mail ist englisch, aber ich hab nit angefangen *g</p><div class="gmail_quote">
On Wed, Feb 11, 2009 at 2:00 PM, Jan Sawicki <span dir="ltr"><<a href="mailto:apsgmbh@gmail.com" target="_blank">apsgmbh@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Let me know if you want to attend !!!!!<br><br><br><br>Discussed will be:<br><br>Short description with live examples of setup we plan for freebone nodes.</blockquote>
<div>in first step i wanna test some more on ap-multiple client, singlehop wds, multihop wds-mesh performance outdoors,.. and if we wann make ptp, we can think of an ntstrme2 revival *g</div><div>but before testing outdoor on real/new test - links,..</div>
<div>i wanna test some stuff indoors tomorrow </div><div>mostly vaps or vlans via wireless, some wds interop issues</div><div>and of course plan the test setup (which links nodes)</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br><br>1. direct EoIP (ethernet over IP) tunnels from freebone nodes to uplinks (n tunnels for n uplinks ?)</blockquote><div>or we will use openvpn tunnel on the routerboards, and can connect them without problems to any other OS,.. </div>
<div>but we also have to test openvpn implementations of routerboards before doing so in general</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
<br>2. OSPF routing between freebone nodes so ^^ tunnels will always find they way to go.</blockquote><div>and experimtental Layer2 (link quality based) routing on different vlans,.. and we will see which one wins in long term monitoring,..</div>
<div>i assume OSPF to be more stable in general, but this comes with a PRICE!!<br><br></div><div>because OSPF needs perfect monitoring of links, as we have static link costs in OSPF</div><div>but as long as no links has some packetloss (which OSPF is not really aware of), or a link gets incredible slow<br>
<br> OSPF will do very fine,..<br><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">3. links monitoring and SMS/email notification of failure<br><br>4. guidelines for building freebone nodes (aluboxes, good quality antennas/cables, poe, tape sealed connectors, maybe small UPS'es(buffor netzteil) for really important nodes)</blockquote>
<div>add also remote power cycle for important nodes </div><div> </div><div>and maybe minipci shielding )-; *g</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
<br>advantages:<br><br><br>- perfect intergration with existing OLSR network, simplicity</blockquote><div>nothing is perfect *g *g *g<br><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br><br>- flexibility and not much overhead</blockquote><div>unfortunately EOIP has significant overhead, and openvpn will not be better,..<br>but we have some 800Mhz boards, and if the tunnel-concentrators are no routerboards it will be ok (if tunnels are going over multiple hops)<br>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br><br>- setup proven to work in large networks<br><br>disadvantages:<br><br>- not open source <br>
<br>- we have to pay for x86 routeros license on uplinks </blockquote><div>but WHO would pay this?</div><div>so i think we will use openvpn instead, or if we are really crazy some "funny" EOIP workaraound</div>
<div>lg Markus</div></div>