Feature #4816
closedDisplay FIB nexthop endpoint ID in status dataset
Added by Junxiao Shi almost 6 years ago. Updated over 5 years ago.
Description
In FibMgmt, include nexthop record’s endpoint ID in status dataset.
This issue includes protocol update and implementation in ndn-cxx, nfd, nfdc, XSLT.
Updated by Junxiao Shi almost 6 years ago
- Related to Feature #4284: FIB EndpointId column added
Updated by Ju Pan over 5 years ago
- Status changed from New to In Progress
- Assignee set to Ju Pan
- Estimated time set to 6.00 h
Updated by Ju Pan over 5 years ago
I knew nothing about NFD mgmt before, I am reading NFD developer guide "Management" section, NFD Management Protocol on Wiki page, and the source code. Are there any other materials I could refer to? Thanks!
Updated by Ju Pan over 5 years ago
Just want to make sure I am on the correct direction before changing the code.
To "include nexthop record’s endpoint ID in status dataset", I need to:
- In ndn-cxx/ndn-cxx/mgmt/nfd/fib-entry.hpp, add a "m_endpointID" private field and corresponding get and set methond. And modify the fib-entry.cpp file.
- In NFD/daemon/mgmt/fib-manager.cpp, at line 128, add "setEndpointID(...)"
Updated by Davide Pesavento over 5 years ago
- Start date deleted (
01/25/2019)
Yes, that seems correct to me. Don't forget to update the protocol in FibMgmt.
Updated by Ju Pan over 5 years ago
Two equal NextHopRecords should have the same EndpointId, right?
Updated by Ju Pan over 5 years ago
- Status changed from In Progress to Code review
- % Done changed from 0 to 80
Updated by Junxiao Shi over 5 years ago
FibMgmt rev41 assigns 0x71 to EndpointId
in dataset. Any reason not to use a number in application range, i.e. 0x80 or above?
Updated by Ju Pan over 5 years ago
No, there is no reason. I don't know there is a range requirement. Do I need to reassign a value?
Updated by Davide Pesavento over 5 years ago
Junxiao Shi wrote:
FibMgmt rev41 assigns 0x71 to
EndpointId
in dataset. Any reason not to use a number in application range, i.e. 0x80 or above?
A lot of type assignments in NFD mgmt protocol already use numbers below 0x80, I don't see the problem.
Updated by Junxiao Shi over 5 years ago
A lot of type assignments in NFD mgmt protocol already use numbers below 0x80, I don't see the problem.
So far, only TLV-TYPEs in ControlCommand use [101, 127] range. Otherwise, they use application range.
https://gerrit.named-data.net/5463 patchset4 displays FaceId+EndpointId as:
faceid=262 (endpointId=1, cost=9)
This is not quite right, because FaceId and EndpointId together identify a nexthop, while cost is an attribute of the nexthop.
The syntax should be:
face=262~1 (cost=9)
Updated by Ju Pan over 5 years ago
The syntax should be:
face=262~1 (cost=9)
Should we omit the EndpointID if it's 0, like
face=262 (cost=9)
Updated by Junxiao Shi over 5 years ago
Should we omit the EndpointID if it's 0, like
face=262 (cost=9)
Yes, that's fine.
Updated by Davide Pesavento over 5 years ago
- Target version set to v0.7
- % Done changed from 80 to 100
Updated by Davide Pesavento over 5 years ago
- Related to deleted (Feature #4284: FIB EndpointId column)
Updated by Davide Pesavento over 5 years ago
- Blocked by Feature #4284: FIB EndpointId column added
Updated by Davide Pesavento over 5 years ago
These changes have been reverted, see #4973.