2011/5/8 Marek Lindner lindner_marek@yahoo.de:
Hi,
- The concept of attaching all hops and their information to the OGM
brings us dangerously close to the problems other routing protocols suffer from (especially Link State): The more a node is away from the source the more its information are outdated. Imagine a 10 hop path - how can we rely on this information at the end of that path ? Furthermore, by making a decision at the end of the path it does not mean the next hop (which has newer information) agrees with us on that path. It might send it the packet somewhere else. All that depends at which point the bandwidth influences the routing decision. Do you mind providing a more precise example how the route switching happens ? I hope you thought about loops..
It is not a real Link State routing, the document maybe is not completely clear. Every node before the OGM rebroadcast, attach its current best path toward the originator node. I could do this without adding an hop list, but simply substituting the TQ with PCE. This modification, as explained in the report, has been done to permit future addition of further information about link such as, for example, the load.
No, it is not exactly like Link State but it goes into the same direction. The problem is this: The more information you propagate through the mesh the more you have to pay attention to not make decisions based on outdated data. In a 4 node testbed you will not see this but as your mesh grows you will. Your document explains how the PCE is calculated but I could not find the section which explains how every individual node comes to a routing decision. That part is essential for the routing loops.
You are perfectly right. I will revise the report as soon as possible to add this information. Certainly routing loops are the worse problem that can occur, so I will think deeply on it. Thanks for the suggestion!
In addition you gain no value by sending all these information. Even if every node had up-to-date information about all other nodes in the mesh and made a decision based on that, the next hop can come to a completely different decision and sends the packet somewhere else. That is one of the main reason why I don't see the multipath routing to work as you envision in the paper.
mmh..if I have up-to-date information, maybe using the OGM sequence number I can tell my neighbor use a specific path, in particular the best path that there was when that specific OGM traversed the network. And this can be valid also for an hypothetical multi-path routing.
Now that I think on it, with this trick can be demonstrated also the absence of routing loops because in every moment I know the path that a packet will follow.
- Bit-rates are not a fixed value either - mistrel for example maintains
a table of 3 different bit-rates. How do you obtain one value ? What about the other bit rate algorithms ?
As far as I know, the value given from driver is the current "best-throughput" one in the minstrel case. In any case is reported the most used by the wireless card.
Ok, your results are not accurate then. Not sure whether this is a problem or not.
- Bit-rates are a best guess too which might drop rapidly as soon as you
try to actually use the link.
For the minstrel case, no estimation of bit-rate is done, the bit-rate is obtained only measuring real traffic. For the other algorithms I have surely to do a more accurate research.
What makes you think this is not an "estimation" ? On wireless you never know the actual throughput unless you use the link. Even then the outcome is highly variable depending on whether you have neighboring nodes that also send traffic, whether there is a sudden interference (microwave is turned on) and so forth and so on.
As answered to Andrew Lunn, I can exploit the bit-rate variability that summarize many aspects of a link (load, node distance, microwave oven, ..). So we can carefully, use this info, opportunely averaged over time to obtain a better description of a link.
In addition I assume that links are used periodically, otherwise why is routing necessary? :)
You assume a steady flow of data which rarely is the case. Imagine a user surfing the web - you'll have peaks of traffic and then longer periods of no traffic. During the peak time the bandwidth estimate will drop and recover quickly as soon as the link is not used anymore (I have seen that many times).
As said before, this is not necessarily a defect, is a better understanding of link conditions. When load is heavy in a sector of the network, packet are routed in another direction as actually happens also with current batman (and this is much valuable!!!!!!)
Cheers, Marek