(this is the view from my former office! Click on it for the full size image)

Short CV

I am a former associate professor at the Dipartimento di Ingegneria dell'Informazione of the Università di Pisa, Italy. I left Unipi in 2021, and now work with Google, where I started in late 2016 after going on leave from Unipi.

My research is in the area of computer networks, most recently on fast packet processing, packet scheduling, network emulation, disk scheduling.

In previous years i did some work on mostly on multicast, erasure coding, and multicast congestion control. Some of this work has been done with colleagues in the RMRG, funded by Cisco and Microsoft Research. More on my research page (or you can look me up on Google Scholar).

I have been a visiting scientist at ICSI/UC Berkeley in fall 2000, 2001, 2002, 2013, 2014. In fall 2003 and 2004 I have worked with the Intel Research Lab in Cambridge (UK) on a project for the Continuous Monitoring of IP networks (CoMo). My collaboration with Intel has continued in fall 2010 at the Intel Berkeley research Lab.

Recently I have worked in the Onelab, Onelab2 and OpenLab EC projects, extending the emulation features of Planetlab. In another recent project, CHANGE, we have worked on defining a Flow Processing platform for the Internet, and this has resulted in the Netmap framework. Previous projects include E-NEXT and COST 264.

I was the General Chair of the Sigcomm 2006 conference in Pisa in September 2006, TPC Co-Chair for Sigcomm 2009 in Barcelona, and CoNEXT 2014 in Sydney, co-editor of a special issue of JSAC on multicast, and co-chair for NGC'99. I have also served as a TPC member for a number of conferences (SIGCOMM, CoNext, EuroSys, NGC, ICNP, Networking, Usenix ATC).

Over time, I ended up writing a few pieces of software, related (and sometimes not) to research and teaching, see the entries on the left column for some references. Some of these things (e.g. dummynet andnetmapFreeBSD. Some of my code apparently has also made it into Windows XP (which incidentally is the reason I had to change my email address!).

In the past I have also contributed a fair amount of code to an open source PBX called Asterisk.

Why I had to change my email address ?

My email address happens to be in the RELNOTES.HTM file which is part of the Windows XP distribution, because of some software that i wrote and that apparently Microsoft used.

Because of this, since early 2002 I (and a few other lucky people) started receiving a lot of email, both direct spam/viruses and backscatter from mail programs and virus filters which fail to detect that the sender's address is forged. I got used to that, and with [quite] a bit of manual deletion and later the help of a few simple patterns in my mail reader's configuration file, i was able to cope with the moderate (some 200 messages a day -- did I say "a lot" earlier ? well, opinions change, read further) amount of spam I was receiving.

Unfortunately, on Aug.19th, 2003 the situation got out of control. The Sobig.F virus hit a number of machines on the net, and I started receiving tens of thousands of emails per day -- 35k on the 19th, over 50k on the 20th and 21th, until the syadmins on my website decided that this was too much load for their boxes (antiviruses have a lot of processing to do) and had to disable the old address.

Eventually I ended up using some bayesian email filter and pushed the processing of my email to a separate box, so I could re-enable the old email address, just in case some old friend tries to contact me. But filters are not perfect, and some false positives are unavoidable, occasionally causing mail to be lost. Unfortunately I cannot avoid using them -- even now, after nine months, I am still receiving between 20k and 50k messages per day (slightly more when a new virus strikes, slightly less when it settles down).