IETF92 was held in Dallas, TX, USA during the period March 22-27, 2015. This is a quick summery of some of the working groups that met during this week.
NETCONF Data Modeling Language (NETMOD)
This working group is working on YANG and YANG models. There was a very large attendance during this meeting as YANG is quickly gaining in popularity. Google presented one draft about how to organize monitoring data in YANG. During the presentation they said that Google would prefer to use NETCONF for both configuration and monitoring. Many of the big customers of networking equipment, eg. Google, Amazon etc., want YANG modules for “everything”. The big vendors like Cisco and Juniper don’t have time to wait for the IETF to standardize everything.
The working group is working on YANG version 1.1. It is getting close to being finished, but there are still a few open issues that needs more discussion before they can be closed. On topic that they have been working on for a long time is how to handle different versions of groups and typedef. This is still not fully resolved.
Network configuration (NETCONF)
The two main topics discussed at this meeting was RESTCONF and call home functionality for both NETCONF and RESTCONF. RESTCONF is a new protocol being standardized by the NETCONF working group and is an HTTP-based protocol that provides a programmatic interface for accessing data defined in YANG, using the datastores defined in NETCONF. RESTCONF is getting close to being finished. There were some discussions regarding the scope of RESTCONF. Some people want to add as many features as possible which mean that the resulting protocol will be more like NETCONF over HTTP than a simple REST API to access YANG data. The majority wants to keep it simple for now and maybe add more features to a future version.
The call home draft adds a feature to both NETCONF and RESTCONF where the managed device can initiate the connection to the manager. This is used when the managed device is behind firewalls or NAT. After the initial connection, the manager will assume its normal role.
DDoS Open Threat Signaling (DOTS) BOF
New BOF to form a working group for standardizing a protocol for signaling upstream about security attacks like DDoS. There already exists equipment for this where big customers of an ISP can install a box that detects a DDoS attack and then signals a box in the ISPs network so that it can automatically mitigate the attack. The problem is that these boxes can only talk to each other as long as they are from the same vendor.
The work will define what a DDoS attack is and how it can be detected. It will then define a protocol to send information about the attack. The suggestion at the BOF was to use IPFIX with some new attributes.
It looks like there is enough interest in this work to form a working group at the next IETF.
Large-Scale Measurement of Broadband Performance (LMAP)
LMAP works on a protocol for doing large scale active and passive measurements. The information model draft is now reasonable stable and it was decided to focus on the data model and implementation to see if that would find new weaknesses in the information model. The basic idea in LMAP is to have a number of Measurement Agents (MA) that are controlled by a Controller. An MA does the actual measurements and when needed it can use a Measurement Peer(MP) to conduct the measurements. Measurement results are sent to a Collector.
YANG and RESTCONF was selected for the data model and control protocol. The upcoming call home functionality of YANG will be used.
Operations and Management Area Working Group, combined with OPSAREA
Have set up a YANG Model Coordination Group. Four people are part of it and each of them will use 1/3 of their time to help write and coordinate all YANG modules within the IETF. Main focus is to make sure that all the IETF YANG modules can work together.
Someone had made a script that automatically pulled YANG modules from active Internet drafts. At the meeting, 113 YANG modules were found.