Ian Clarke wrote:
> On 30/11/05, Gordan Bobic <gordan at bobich.net> wrote:
> 
>>Matthew Toseland wrote:
>>
>>>Umm, please read the presentation on 0.7. Specializations are simply
>>>fixed numbers in 0.7.  The problem with probabilistic caching according
>>>to specialization is that we need to deal with both very small networks
>>>and very large networks.  How do we sort this out?
>>
>>It's quite simple - on smaller networks, the specialisation of the node
>>will be wider. You use a mean and standard deviation of the current
>>store distribution. If the standard deviation is large, you make it more
>>likely to cache things further away.
> 
> 
> You are proposing a fix to a problem before we have even determined
> whether a problem exists.  I am not currently aware of any evidence
> that simple LRU provides inadequate specialization, or that we need to
> enforce specialization in this way.
> 
> In other words: If its not broken, don't fix it (words every software
> engineer should live by).

Having just put two nodes up, one with unlimited bandwidth (well, 
100Mb/s) one with less, and seeing both of them sit at the maximum 
bandwidth set or maximum CPU usage, whichever runs out first, tells me 
that there likely is a problem.

It seems obvious to me that without specialisation there can be no 
routing other than random/flooding - and I am not seeing particularly 
pronounced specialisation. The only reason it _seems_ to work is because 
popular content gets caches on most nodes.

Gordan

Reply via email to