> looks like I am wrong. Someone dug into the code and found its the node
> LRUed by last use time that is dropped. This normally is one with a very low
> CP... This means we are only using CP to sort the list which is harmless.
With NGR there is no reason that we even need the concept of a CP
On July 23, 2003 09:08 pm, Ian Clarke wrote:
> On Wed, Jul 23, 2003 at 09:02:49PM -0400, Ed Tomlinson wrote:
> > On July 22, 2003 02:38 pm, Toad wrote:
> > > > This is basicily what the current code tries to do. Quite probably I
> > > > should make nodes that do not have connection have a very low
On Wed, Jul 23, 2003 at 09:02:49PM -0400, Ed Tomlinson wrote:
> On July 22, 2003 02:38 pm, Toad wrote:
>
> > > This is basicily what the current code tries to do. Quite probably I
> > > should make nodes that do not have connection have a very low CP. Note
> > > about the only thing CP is used f
On July 22, 2003 02:38 pm, Toad wrote:
> > This is basicily what the current code tries to do. Quite probably I
> > should make nodes that do not have connection have a very low CP. Note
> > about the only thing CP is used for the NG code to to drop nodes (and
> > give the sort order for the dis
Andrew Rodland <[EMAIL PROTECTED]> writes:
> IIRC, since the "backoff" deallie went away, the criterion for
> dropping nodes is their CP going below minCP (0.01 by default).
>
> --hobbs
>
I looked around a lot in the source, and while there's code indicative
of this kind of behavior in the past
Ian Clarke <[EMAIL PROTECTED]> writes:
> On Tue, Jul 22, 2003 at 01:52:04PM -0500, Edgar Friendly wrote:
> > > Drop nodes? We don't drop nodes based on CP, do we?
> > >
> > IIRC, the node in the RT with the lowest CP is dropped to make room
> > for a new node.
>
> I hope not, references are supp
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Ian Clarke wrote:
|>IIRC, the node in the RT with the lowest CP is dropped to make room
|>for a new node.
|
|
|I hope not, references are supposed to be dropped on an LRU basis, not
|on the basis of their CP. If someone change that then it is a miracl
On Tue, Jul 22, 2003 at 01:52:04PM -0500, Edgar Friendly wrote:
> > Drop nodes? We don't drop nodes based on CP, do we?
> >
> IIRC, the node in the RT with the lowest CP is dropped to make room
> for a new node.
I hope not, references are supposed to be dropped on an LRU basis, not
on the basis
Toad <[EMAIL PROTECTED]> writes:
> > This is basicily what the current code tries to do. Quite probably I should
> > make nodes that do not have connection have a very low CP. Note about
> > the only thing CP is used for the NG code to to drop nodes (and give the
> > sort order for the display s
On Mon, Jul 21, 2003 at 08:37:14AM -0400, Ed Tomlinson wrote:
> On July 20, 2003 03:53 pm, Ian Clarke wrote:
> > A thought just occurred to me. It is quite possible, that from the
> > perspective of the NG routing algorithm - that it will be preferable to
> > send a request to a poorly suited node
On July 20, 2003 03:53 pm, Ian Clarke wrote:
> A thought just occurred to me. It is quite possible, that from the
> perspective of the NG routing algorithm - that it will be preferable to
> send a request to a poorly suited node to whom a connection is already
> open, than to a well suited node to
On Monday 21 July 2003 12:04 am, Ian Clarke wrote:
> > Basically taking probabilistic caching one step further. Come up with a
> > ngrouting algorithm to perdict our incoming data requests, and if the
> > item that is currently being processed is more likely to be requested
> > then our current lea
On Sun, Jul 20, 2003 at 08:45:51PM -0700, Todd Walton wrote:
> > Fortunately, a much easier way would be to simply say that all nodes in
> > the RT must already have connections opened to them the moment they
> > enter the routing table. If such a connection cannot be established,
> > then the nod
On Sun, Jul 20, 2003 at 10:04:22PM -0500, Tom Kaitchuck wrote:
> I don't see how it is a security threat. Ether it will respond to your
> requests or not. If it is trying to monitor your traffic, what difference
> does it matter who connected to who.
Because someone to whom your node deliberatel
On Sun, 20 Jul 2003, Ian Clarke wrote:
> Again, if we were going with my first (undesirable) suggestion, perhaps,
> but my second (desirable) suggestion makes all of this irrelevant.
Your second suggestion:
> Fortunately, a much easier way would be to simply say that all nodes in
> the RT must
On Sunday 20 July 2003 05:07 pm, Ian Clarke wrote:
> > Why not make it go both ways? Make Routing Table == Open Connections -
> > Transients.
>
> Well, I am not necessarily sure that it is wise to route messages to
> nodes that initiated the connection to us since that could pose a
> security threa
On Sun, Jul 20, 2003 at 07:30:00PM -0700, Todd Walton wrote:
> 1) As I understand it so far, when it comes time to hand a request to some
> other node, fred takes a look at the stats it already has on the nodes in
> its routing table and figures out which one would best serve this request.
> Wh
On Sun, 20 Jul 2003, Ian Clarke wrote:
> A thought just occurred to me. It is quite possible, that from the
> perspective of the NG routing algorithm - that it will be preferable to
> send a request to a poorly suited node to whom a connection is already
> open, than to a well suited node to whic
On Sun, Jul 20, 2003 at 04:04:07PM -0500, Tom Kaitchuck wrote:
> > Why is this bad? Well, NG would be making a perfectly sensible and
> > rational decision given that its goal is to minimize the time required
> > to retrieve the data for *this* request, but it doesn't account for the
> > fact that
On Sunday 20 July 2003 02:53 pm, Ian Clarke wrote:
> A thought just occurred to me. It is quite possible, that from the
> perspective of the NG routing algorithm - that it will be preferable to
> send a request to a poorly suited node to whom a connection is already
> open, than to a well suited n
A thought just occurred to me. It is quite possible, that from the
perspective of the NG routing algorithm - that it will be preferable to
send a request to a poorly suited node to whom a connection is already
open, than to a well suited node to which we would need to establish a
connection.
Why
21 matches
Mail list logo
