"Routing Table Size" in Exercise 4d

Moderator: Peer-to-peer und Grid Computing

Benutzeravatar
Pan
Erstie
Erstie
Beiträge: 20
Registriert: 3. Nov 2008 15:33

"Routing Table Size" in Exercise 4d

Beitrag von Pan »

Hi,

I'd like to ask which slide to stick to when calculating the "routing table size".

On Page 25 it says:
Basic Chord keeps track of predecessor and successor
Finger tables not needed for correctness
But on Page 30 it says:
Predecessor links, finger tables – Only needed to improve performance
So it's a np-hard problem to decide whether "predecessor" is necessary for the "minimum size" or not. :roll:

Benutzeravatar
BadTaste
Endlosschleifenbastler
Endlosschleifenbastler
Beiträge: 181
Registriert: 18. Apr 2005 16:40
Wohnort: Darmstadt
Kontaktdaten:

Re: "Routing Table Size" in Exercise 4d

Beitrag von BadTaste »

Please consider the case for an efficient routing procedure.

Juri
Mausschubser
Mausschubser
Beiträge: 64
Registriert: 13. Mär 2007 10:48

Re: "Routing Table Size" in Exercise 4d

Beitrag von Juri »

Hi,

i had another question to this task...

a finger table looks like this as known from the above tasks and slides: (i, targetID, successor), where targedID = Hash(IP_Comp_X) and succ = Hash(IP_Comp_Y)...where do we save 32bit-IP-Addresses and 16-bit-Ports in the finger table?

maybe i missed something...thank you for any help :wink:

Juri
Mausschubser
Mausschubser
Beiträge: 64
Registriert: 13. Mär 2007 10:48

Re: "Routing Table Size" in Exercise 4d

Beitrag von Juri »

Hi,

question solved...for everybody that had the same question see on page 4 of the Paper about Chorde "pdf-file", left side and last paragraph..

kaiserj
Erstie
Erstie
Beiträge: 18
Registriert: 13. Okt 2009 19:55

Re: "Routing Table Size" in Exercise 4d

Beitrag von kaiserj »

hey juri,

von welchem "pdf-file" sprichst du da?

Benutzeravatar
E.d.u.
Nerd
Nerd
Beiträge: 633
Registriert: 3. Feb 2004 17:01
Wohnort: Darmstadt
Kontaktdaten:

Re: "Routing Table Size" in Exercise 4d

Beitrag von E.d.u. »

Denke das hier:
Chord:A scalable peer‐to-­peer lookup service for Internet applications

mister_tt
Kernelcompilierer
Kernelcompilierer
Beiträge: 502
Registriert: 29. Sep 2008 15:54

Re: "Routing Table Size" in Exercise 4d

Beitrag von mister_tt »

Here is the link to the paper: http://pdos.csail.mit.edu/papers/chord: ... igcomm.pdf

The problem is that the paper says that the finger and the routing table is exactly the same thing. It says:
Each node, n, maintains a routing table with (at most) m entries, called the finger table.
The only difference is that the task asks for the size of the complete routing table, but what is the difference? A complete routing table as stated in the task does not make sense at all, right? Because why should I have an entry for every node in the system? If I have that, I do not need Chord... :?:

Benutzeravatar
BadTaste
Endlosschleifenbastler
Endlosschleifenbastler
Beiträge: 181
Registriert: 18. Apr 2005 16:40
Wohnort: Darmstadt
Kontaktdaten:

Re: "Routing Table Size" in Exercise 4d

Beitrag von BadTaste »

Hi,
in the exercise, complete means all information that can be used during routing as discussed in the lecture.
I.e., finger table + ....

mister_tt
Kernelcompilierer
Kernelcompilierer
Beiträge: 502
Registriert: 29. Sep 2008 15:54

Re: "Routing Table Size" in Exercise 4d

Beitrag von mister_tt »

So you mean successor (which is stored in the finger table anyway), predecessor and finger table?

Benutzeravatar
BadTaste
Endlosschleifenbastler
Endlosschleifenbastler
Beiträge: 181
Registriert: 18. Apr 2005 16:40
Wohnort: Darmstadt
Kontaktdaten:

Re: "Routing Table Size" in Exercise 4d

Beitrag von BadTaste »

no further comment :mrgreen:

mister_tt
Kernelcompilierer
Kernelcompilierer
Beiträge: 502
Registriert: 29. Sep 2008 15:54

Re: "Routing Table Size" in Exercise 4d

Beitrag von mister_tt »

:evil:

Antworten

Zurück zu „Peer-to-Peer und Grid Computing“