Posts

I am just returning form the 2nd International Summer School on Network and Service Management. During some SNMP experiments, something interesting happened: A networking device used for the tests which has been silently running for years at the University of Twente suddenly went offline. We later found out that someone in Twente decided that today is the perfect day to move the box to another place…

Later on, during another lab sessions, we experienced a problem where packets originating form the notebooks of the summer school participants experienced high packet loss at the border router of our university, rendering ssh useless. There was no packet loss for ssh traffic originating from other locations so we worked around the problem by letting people ssh to other nodes first before they did connect to our machines. We also had the feeling that the packet loss increased with the number of concurrent ssh sessions originating form the summer school network. Of course, this problem did not happen before and this strange behavior apparently also did not affect any other sites…

The program of the AIMS 2008 conference is online and the registration has opened. The event features next to six technical paper sessions a keynote, six tutorials, a PhD workshop, and a social event for “networking 2.0”. Please join us in July!

Today, I had a closer look at two commercial NETCONF implementations and getting some small examples going was not as easy as I hoped. While I spotted some clear implementation bugs, there are also quite a few things where the specifications (RFC 4741 and RFC 4742) are either somewhat ambiguous or totally silent. It seems some well organized interoperability testing would be helpful to improve implementations and specifications.

I am giving a hands-on introdution to NETCONF and YANG at the 2nd International Summer School on Network and Service Management (ISSNSM). The summer school takes place at the University of Zurich on June 2-6, 2008.

I am just returning from NOMS 2008, which took place this year in Salvador da Bahia in Brazil. While the location is very close to some nice beaches and the location almost guarantees sunshine, it seems that technically interested people end up spending almost a whole week in those stereotype hotel meeting rooms without windows, cooled down to a level that it sometimes feels like entering a refrigerator. I am now wondering if it is possible to exchange scientific ideas effectively at beaches. If the answer would be yes, we should start doing so and stop spending a week in hotel fridges. If the answer is no, we might question the need to have locations with lots of sunshine and beaches for scientific events like NOMS. But anyway, people should really mark MANWEEK 2008 in their calendar; it takes place on the Samos island…

Today, the formation of the NETMOD working group of the IETF has been announced. The group is chartered to develop a standard data modeling language for NETCONF. The starting point for the design is YANG. Overall, it took the YANG design team about half a year from the submission of the YANG specification to the IETF with a request for a BOF until working group formation. While the road was somewhat bumpy, the target has been reached pretty much on time.

The 71st IETF meeting is over and I have recovered from “the IETF cold”. The good news is that the NETCONF data modeling requirements discussion in the IETF is over. There is work going on to draft a charter for a NETCONF data modeling language working group and the YANG specifications are likely becoming the basis of this IETF effort. The details of course need to be further discussed and rough consensus within the IETF has to be reached. But with some luck, we will see a first IETF YANG data modeling language meeting at the 72nd IETF meeting, which takes place in Dublin end of July.

I travelled to the 71st IETF meeting in Philadelphia to do some IETF and NMRG business. What I did not know by the time of arrival in Philadelphia was that I would follow the IETF meeting mostly from my hotel room. But on Monday, it became very clear that I got a bad cold (the IETF cold as some say) and on Tuesday I managed to sleep 15(!) hours in a row. Tomorrow, we have the NMRG meeting and I am looking already forward to sit in the plane to Amsterdam later that day.

I am spending significant time this week in the train. First, I went to the EMANICS P2P workshop hosted by the University of Zurich. From Zurich, I took the train to Brussels in order to attend the EMANICS project review 2.0. While hopping from one EMANICS event to another, I learned that not having Internet is sometimes a real win in getting things done. And finding the right page in Korea can also save you the day.

About a year ago, I posted a note concerning the power consumption of the Internet and the need for a Green Internet. This note triggered some email exchanges and today I did attend a conference organized by some government organizations and BITCOM discussing the resource efficiency of information and communication technology.

While the energy consumption of data centers is usually associated with the energy usage of IT hardware (mainly server systems) and cooling systems, I found it interesting that some people reported a major impact of the software on the energy consumption. Of course, we all know that software has increased its “weight” constantly over the last years because time to market is more important in many cases than designs for efficiency and scalability. But as we move towards Green Software, we might have to rethink the way we do software development and get back to the proper design of truly efficient and scalable solutions.