Differences between revisions 9 and 16 (spanning 7 versions)
Revision 9 as of 2008-12-18 11:56:03
Size: 2611
Editor: alpt
Comment:
Revision 16 as of 2008-12-18 12:08:52
Size: 3107
Editor: alpt
Comment:
Deletions are marked like this. Additions are marked like this.
Line 2: Line 2:
 * accurate study, tests and enhancement of the performance of the Netsukuku
  
addressing system:
 * accurate study, tests and enhancement of the performance of the Netsukuku addressing system:
Line 11: Line 10:
  the success and scalability of the Netsukuku network
 * addressing algorithms that minimise the measurement
  inaccuracies: the hierarchical topology helps to organise a
  the success and scalability of the Netsukuku network.
                See http://netsukuku.freaknet.org/doc/main_doc/topology.pdf
 
* addressing algorithms that minimise the measurement inaccuracies:
                the hierarchical topology helps to organise a
Line 19: Line 19:
  it is possible to develop methods that minimise it   it is possible to develop methods that minimise it.
                See http://netsukuku.freaknet.org/doc/main_doc/topology.pdf
Line 21: Line 22:
   a distributed, non hierarchical and decentralised system of hostname management    a distributed, non hierarchical and decentralised system of hostname management.
                See http://netsukuku.freaknet.org/doc/main_doc/andna.pdf
                http://netsukuku.freaknet.org/doc/main_doc/ntk_rfc/Ntk_p2p_over_ntk.pdf
                and http://dev.hinezumi.org/browser/netsukuku/trunk/pyntk/ntk/core/p2p.py
Line 23: Line 27:
   a node could easily forge an
  
entry in the routing table of a neighbouring node, injecting false information about links in the network. The attack would just create a temporary local damage, thanks to the distributed nature of the QSPN. However the optimal solution is to prevent these attacks using a cryptographic signature scheme
   a node could easily forge an entry in the routing table of a neighbouring node, injecting false information about links in the network. The attack would just create a temporary local damage, thanks to the distributed nature of the QSPN. However the optimal solution is to prevent these attacks using a cryptographic signature scheme. See http://netsukuku.freaknet.org/doc/main_doc/qspn.pdf
Line 26: Line 29:
the  multipath routing is a method which sends packets of a same                 the multipath routing is a method which sends packets of a same
Line 30: Line 33:
  a small load on the interested nodes   a small load on the interested nodes. See http://netsukuku.freaknet.org/doc/main_doc/ntk_rfc/Ntk_caustic_routing
Line 32: Line 35:
 * software implementation of Viphilama, an Internet tunnelling system that connects nodes which aren't physically linked  * software implementation of Viphilama, an Internet tunnelling system that connects nodes which aren't physically linked. See http://netsukuku.freaknet.org/doc/main_doc/ntk_rfc/Ntk_viphilama
Line 35: Line 38:
  nodes which are sharing their connection.   nodes which are sharing their connection. See http://netsukuku.freaknet.org/doc/main_doc/ntk_rfc/Ntk_IGS
Line 40: Line 43:
 * Beta testing

 * Write more documentation
Line 45: Line 44:

 * Draw pictures to explain the basic concepts of the procotol

 * Animation of the various mechanisms of the Npv7

 * Looking for new TODOs, bugs

TODO

  • accurate study, tests and enhancement of the performance of the Netsukuku addressing system:
    • the hierarchical topology of Netsukuku is dependent on the addresses associated to each node. Since the network is dynamic, the addresses can change over time, thus the topology must be updated. When the topology changes, other parts of the network must be updated too, f.e. the routing tables of the affected nodes. Therefore, an efficient minimisation of the number of topology updates is critical for the success and scalability of the Netsukuku network.

      See http://netsukuku.freaknet.org/doc/main_doc/topology.pdf

  • addressing algorithms that minimise the measurement inaccuracies:
    • the hierarchical topology helps to organise a large number of nodes, however this structure hides some information about the routes; as a consequence, the routing table known by a node is an approximation of the best absolute routing table available to it. The approximation is dependent on the particular addresses assigned to each node, therefore it is possible to develop methods that minimise it.

      See http://netsukuku.freaknet.org/doc/main_doc/topology.pdf

  • software implementation of ANDNA:
  • secure QSPN:
    • a node could easily forge an entry in the routing table of a neighbouring node, injecting false information about links in the network. The attack would just create a temporary local damage, thanks to the distributed nature of the QSPN. However the optimal solution is to prevent these attacks using a cryptographic signature scheme. See http://netsukuku.freaknet.org/doc/main_doc/qspn.pdf

  • improvement, test and implementation of the Caustic Routing:
  • stable implementation of the Netsukuku software
  • software implementation of Viphilama, an Internet tunnelling system that connects nodes which aren't physically linked. See http://netsukuku.freaknet.org/doc/main_doc/ntk_rfc/Ntk_viphilama

  • software implementation of a system of distributed Internet connections through Netsukuku:
  • experimental tests
  • Ultimate the incomplete RFCs: Netsukuku_RFC

  • Translate the documentation

Ntk_TODO (last edited 2009-09-18 18:23:05 by alpt)