The Moment (Mobility Management and Networking ) Lab is an experience in itself and the weeks leading up to what we researched this summer were certainly amazing!!!
My first week in Santa Barbara was mostly settling into my new house and work. I spent most of my weekend exploring the local sights here. Santa Barbara is located in between the ocean and the mountains and is hence so scenic and beautiful. The weather here is forever pleasant and relaxing. Also, the beaches are right around the corner and you will find them full of students on the weekends. As far as work concerns, Prof. Elizabeth Belding had absolutely everything ready for us when we got here, from a Santa Barbara welcome package to our laptops and lab accounts. My Student mentor Amit Jardosh (PhD Candidate) was extremely helpful in getting me started on some reading material on the project etc. Overall, between work and home, there never was a dull moment.
My second week in UCSB was more productive in terms of the amount of work completed. I started the week with a crash course on wireless networking by prof. Belding and reading a few papers on topics such as Multimedia Traffic Performance on Mesh Networks and Understanding Link Layer behavior on a congested network. Amit and prof. Belding took some time out explaining to me the wireless mesh network that is set up in the Harold Frank building of UCSB. On Tuesday, I started some actual work on my project which involved setting up a communication application on two laptops. The goal was to be able to communicate through voice transmission in between the laptops. For this, we installed the RAT (Robust Audio Tool) software on the machine and had it ready to communicate.
This weekend I visited the Santa Barbara downtown area and the pier etc. The local outdoor mall had a very open nad fun atmosphere! As far as research goes, this week was very eventful. My project took a slight turn which I guess is very natural during research. I realised that RAT being a purely audio tool does not support any session initiating or call registering features. And, the Asterics PBX can only provide service for audio devices with the SIP or VoIP protocols. Hence, we decided to write our own low-key session initiation protocol using the Perl scripting language. Perl is a new language for me and hence I had to study it before writing the Socket program. The basic functionality of the program was to establish a connection between the client (machine placing the call) and server (gateway), and request the server for the IP address of the computer we wish to call. The program will then initiate RAT with the provided IP address.
It seems like this week just flew by! Raji, another student in my MOMENT lab, and I visited the Santa Barbara Zoo this weekend. It was small but so beautiful and all the animals were just great especially the Giraffe!. In the lab, I got further with my work. In order to transmit voice packets uniformly through RAT, I intended to install a media tool like VLC player. Unfortunately, the VLC player just wouldn't work so I switched to using an audio CD to transmit audio files between the two client computers. The next step was to install tethereal, a network protocol analyser to analyse the voice packets sent and received. I was able to record some of the packets transmitted across the machines in a tcpdump file and obtain some of the important fields from the file using pre-written perl scripts. The following week, I should be able to work more on these files and further my research progress.
This week, I was able to work more on my tcp dump files. Using tethereal, I was able to filter out the packets that were relevant to my experiment. (The packets being sent between the two machines using the RAT application). After saving the data in a more readable format in a different file, the next stage was to analyse these packets. The various metrics with which these files could be analysed include delay, inter packet delay and missing packets. On a more fun note, the seventh book in the Harry Potter series was released this Saturday and much of my weekend was spent reading it!
In order to analyse my files (with data retreived from the tcp dump files), I needed to code some perl scripts that will help determine some metrics such as delay, missing packets etc. For this, I decided to write two scripts, the first script would read in the files and calculate the inter packet delay in the transmission. It will also store the ip.id and timestamp information of each packet ontoa separate output file. The second perl script would open the output files from the sender and receiver side and calculate the end-to-end delay and the total number of missing packets. Since perl is still a relatively new language to me, coding the 2 scripts was a very learning experience for me. As of this friday, I completed the 1st script and am still debugging my second script. This weekend was as fun as any other. I went to a nice dinner and coffee with some friends from my lab and basically enjoyed a relaxing weekend!
With my two scripts to obtain the metrics ready, I proceeded to conduct my experiment on the wireless network. This was not a very smooth process and I encountered several difficulties before performing my experiment. First of all, I had to edit my scripts to reflect the new IP addresses of my machines on the wireless network. Secondly, with the help of my student mentor Amit, I had to register my two machines onto the UCSB MeshNet. While conducting my experiment, I had some trouble with the RAT tool. It did not seem to work on the wireless network as it did on the wired network. The rest of my week was spent trying to solve the problem with RAT and getting my first experiment on wireless to work. As for the weekend, some friends and I went to downtown Santa Barbara for the Fiesta. We had a great time throwing confetti all over and pigging out on some excellent mexican food!
Continuing on from the previous week, I worked on getting my experiment to work on the wireless mesh network. In the process, we decided that the best approach was to take three sets of readings with different number of hops (between the 2 laptops communicating) to obtain a cummulative view of how the packets are being transmitted over the network and the different factors that come into play. While setting up my experiment on 2 hops, I further edited my scripts to filter my data more efficiently and gather information on the inter-packet delay, end-to-end delay and number of missing packets. My aim was to be able to plot some graphs of delay vs timestamp, number of hops vs timestamp etc to observe any patterns on the communication process. This weekend, I visited some of my friends in LA and had a tour of Holywood Blvd, Beverly Hills and some of the popular tourist attractions there. It sure was difficult getting back to reality after visiting the surreal world of the rich and famous.
This week, with the help of my mentor, I started performing my experiments on different hops. This required creating paths any data should take while being sent and received by the two laptops. It was decided that I would do the experiments on 4 different hops. Due to various random factors that could affect an experiment, I took atleast two sets of readings at the same hop. Also, for the purpose of having a more readable graphs, I created scripts that would alter the timestamp, ip.id etc into a more readable unit. As for the weekend, this being my last one in Santa Barbara, I decided to visit downtown one last time with my friends! We had a great time just walking down state street and to the beach simply hanging out!
This being my final week in Santa Barbara, I was determined to have a blast as well as round up my work. During the weekend, my friends and I visited downtown again and spent the whole day hanging out on State street. As for lab, this week sure was hectic. I had to write my report and finish up a final run of my all my experiments. My graphs looked good and I was very satisfied with my work!On that note, I have to thank my friend Lara who really helped me out with my report. Without her all the linux odt to microsoft word to pdf conversions of my report would have driven me crazy! Also, on Friday, everybody in our Moment lab is planning to have a little party celebrating the end of a good summer and the defence of one of the phD students here! Although the summer has been absolutely great, I think I'm ready to get back to college station and start the next semester! Viva La Santa Barbara!!!