P2PSIP diagnostics

1
P2PSIP diagnostics

• draft-zheng-p2psip-diagnose-03

2
Differences from last version

• Consistent with p2psip base protocol
• Split the Echo method functions into Ping,
  Route_Query and new Echo
• Define new Error Codes for connectivity check

3
Diagnostic scenarios

• Diagnostics built into the routing protocol
• E.g. Keep-alive between neighbors
• Diagnostics to monitor the overall health of the
  network
• Overall resource consumption, malicious nodes,
  etc
• A diagnostics server may be needed
• Diagnostics for a single node
• The quality of the connectivity to the network
• Report where the problem happens when the source
  cant reach its destination

4
Diagnostics Information

• Some of the information may be included in the
  Error Response
• Connectivity reachability information
• Misrouting information
• Loop detection information
• Timestamp HopCounter
• Topology information(number of entries in the
  neighbor table, etc.)
• Resource information
• Congestion status information
• Response peer information

5
Ping(extended)

• Test the connectivity along a path
• Measure the connectivity quality along the path
• Measure the overlay hops to the destination ID
• Measure the RTT to the destination ID

6
new Error Codes

• Need Error Codes to indicate the following types
  of error
• Underlay Destination Unreachable
• Underlay Time exceeded
• Message Expired
• Upstream Misrouting
• Loop detected
• TTL hops exceeded
• Can be as a response to Ping and other messages
  when failures occur

7
Example(Ping Error)

Ping
N40
N60
Error(Loop detected)
Ping
P2p overlay
N80
N20
8
Route_Query(extended)

• Retrieve diagnostic information(iterative mode)

N40
N60
Route_Query
P2p overlay
Route_Query
N80
Route_Query
N20
Destination
9
Echo

Echo
N40
N60
Echo Response
Echo Response
Echo
Echo
P2p overlay
N80
N20
Echo Response
10
Pros and Cons

• Echo is efficient, but may introduce potential
  DoS attack
• Iterative Route_Query mitigates the security
  threat, but introduce more messages to the
  overlay
• Use them according to the deploy scenarios
  specifically

11
Authorization

• Authorization is important in retrieving the
  diagnostic information

12
To be added

• Diagnostics server usage

13
Open issues

• Should we extend Ping.
• Should we extend Route_Query.

14
Adopt?

• Interesting and a good start for the p2psip
  diagnostics?
• If so, shall we adopt the draft as a WG item to
  move the work forward?