On Sat, Jul 2, 2011 at 10:43 AM, Matthew Toseland <toad at amphibian.dyndns.org > wrote:
> I don't think I fully understand load management, and I would appreciate > help from people who do! Unfortunately our theorists are mostly interested > in routing, and mostly tend to go away after a while anyway. It is also > possible that the problems on testnet are caused by simple bugs, I have > solved many but more must remain. > I think you are diving straight into the details of a solution, which makes it very difficult for anyone to really give you useful feedback. We don't understand the goals, or how you arrived at this solution beginning from these goals. Consequently I think we need to take a step back and ask ourselves a few fundamental questions, in particular: What would the perfect load-management system do? In other words, what are our goals? The challenge is that we have a scarce resource, which is a node's ability to process and forward requests and responses, and we wish to limit network load such that this scarce resource isn't exhausted, and is fairly allocated between requestors. The extent to which this scarce resource is used, and where it is used, is determined by the rate at which nodes spawn requests, the type of those requests, and how those requests are forwarded. Perhaps if you can describe your proposal by starting with a goal, and then explaining how this proposal achieves this goal, it will help us all to think about it and understand what you are trying to achieve. Ian. -- Ian Clarke Founder, The Freenet Project Email: ian at freenetproject.org -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20110702/cc935b5b/attachment.html>
