P. Oscar Boykin

I tend to think of myself moderate to left leaning. I don’t like to see people become cheerleaders for a particular political party. Too many people treat politics like sports and pick a team to pull for, and they pull for that team no matter what. Another end of the spectrum is to say “that the similarities between the two parties tower over the dwindling differences” implying that it does not matter which of those wins. At least there is a major difference between how the democrats and the republicans are funded.

The US Census Bureau tracks voting in the US. In 2000, 60% of citizens and 86% of registered voters turned out to vote (interestingly, women have had better turnout rates than men since 1980). In that same election cycle, of those registered, 44% were Democrats, 33% were Republicans and 23% were independents or registered with minor parties. This also means that there were 33% more registered Democrats than Republicans (but as the article mentions, independents always decide the outcome).

Despite the fact that there were 33% more registered Democrats, OpenSecrets tells us that Republicans raised 38% more money! Democrats raised $8.25 per registered Democrat, while Republicans raised $15.23 per registered Republican. Even if Republicans are not as successful registering citizen voters, they can be more effective in getting out their message due to raising more money in total, and almost twice as much per voter. Where does this extra money come from? Of the top 10 contributors to the Democratic Party, 7 were labor unions, 2 were corporations, and 1 was a PAC. Of the top 10 contributors to the Republican Party, 8 were corporations, 1 was an activism organization, and 1 was a PAC.

This funding situation can be improved. Republican 2000 presidential hopeful Alan Keyes had one thing right: “no dollar vote without a ballot vote”. I would go even further, both Arizona and Maine have full publicly funded elections. This makes sure that voters are funding elections, and not large corporations or special interests. If you have ever wondered why John McCain is so independently minded (and spearheaded federal campaign finance reform), it might be that he is elected through publicly funded elections.

I am really getting off the tracks here, but the point I wanted to make is that there are real and major differences in the way that the Democrats and Republicans are funded. OpenSecrets can help you “follow the money”. Also, FundRace is an interesting site to check to see how geography and funding are correlated. Public Campaign and Fair Elections are working to change the way elections are funded. Lastly, the Center for Voting and Democracy is working to improve our voting systems, including advocating a greatly improved system called Instant Runoff Voting.

I just got back from a visit to Gainesville, FL. I spent most of my time meeting with my real estate agent, Mika Harris. I think I have found a house, but I am yet to close, so one never knows. It was nice to compare the real estate market in Gainesville to Los Angeles; it turns out that LA is a bit more expensive.

Recently, the Brunet project moved our version control system from CVS to Gnu Arch. Arch is a very powerful system. We are using the Gatekeeper Model, where each of our developers has their own Arch repository (mine is here), and Kamen M. Lozev will merge everything into the mainline tree. I am really impressed with Arch. Coming from CVS it seems a bit confusing at first but that is because it can do things that CVS simply can’t do (like star-merge). Once you see how much easier it makes group development, I think you’ll never look back.

I saw The Corporation last night. I think it is a must see. I tend to feel discouraged by such films, but there are some things I can do. I can buy sweatshop-free clothes from NoSweat, I can invest my money in Green Century Funds, I can use environmentally safe cleaning products. In my opinion, people need to vote politically, as well as with their purchases. Lastly, I think improving voting systems (such as instant run-off voting, advocated by Santa Monica Ranked Voting) could help get more voices heard in the US political arena.

UCLA has wireless lan access for BOL. They provide software to access the vpn, but this software is not Free Software. There is a Free Software client called vpnc. This software is also packaged for debian. So, I set out to install it.

Install was easy (apt-get install vpnc). The configuration was a bit confusing because UCLA does not give the explicit information you need to configure the software. It works this way: there are several configuration parameters which they do not mention on their configuration pages (for instance here). Instead, when you download their software, it comes with a configuration file embedded. They should just give all that information on their website. They didn’t so here it is. There are two things they do not mention on their site, the GroupName and the GroupPwd. These can be found by downloading the closed source linux client, and looking for a file called UCLAVPN.pcf in that archive. This file tells us the following:

GroupName=UCLAIPSEC
enc_GroupPwd=55E136A488F994A7087641D7270E734F33FC67CF45EFF7BE44B2C16ACBACCE48B0962ADD98CB90CBA987050D46C7D849FAAED76E481A269C

Now, the password is not really encrypted, but what I call scrambled (which means it is broken encryption that anyone can decrypt). Cisco knows that this is insecure. UCLA has inconvienced us for “security”, but then used a totally insecure system (great huh?) The vpnc author has a web page that can unscramble the password. That’s all the information we need, now we just set up vpnc.

Write the following:

IPSec gateway vpn.ucla.edu
IPSec ID UCLAIPSEC
IPSec secret <unscrambled password>
Xauth username <your bol id>
Xauth password <your bol password>

to the file /etc/vpnc/default.conf. If you do not have a BOL account, you can create a BOL account on the web. That’s it.

What should happen when you bring up your wireless card, it will create an interface (say eth1) with an ip address. This looks good, but you cannot connect anywhere. Then you type “vpnc-connect” (as root), and you get a second interface (called tun0), which is the IPSec tunnel. Then all your traffic is encrypted and routed over that tunnel. Works for me in 7440 Boelter and all the way to Professor Roychowdhury’s office as well.

Good luck!

The DSN 2004 conference is just about over. It was my first year in attendance and I enjoyed it. I gave a talk entitled “Fault Tolerant Computation on Ensemble Quantum Computers” (pdf of the paper). Since the DSN audience does not have many quantum specialists and the talks were limited to 25 minutes, this talk, unfortunately, does not have much time to get into the real issues we address. I did see many talks that were related to my interest in ad-hoc networking.

Firstly, there was Performance and Dependability of structured peer-to-peer overlays. In this paper they use several P2P traces to evalutate the overhead of Pastry. The traces record the the time nodes join and leave the system. I am interested in using these same traces to drive some of our simultations. The traces come from three papers, A Measurement Study of Peer-to-Peer File Sharing Systems, Understanding Availability (trace data online), and Feasibility of a Serverless Distributed File System Deployed on an Existing Set of Desktop PCs.

Secondly, there was a nice talk on Robust Aggregation Protocols for Large-Scale Overlay Networks. We have considered such algorithms for Brunet. It was good to see some analysis on exactly these topics. This work is a part of their Bison project.

Thirdly, there were two nice, practically oriented papers on coding. One on A Practical Analysis of Low-Density-Parity-Check Erasure Codes for Wide-Area Storage Applications was very interesting. They looked at the efficiency of various codes for finite block size. They showed a nice plot of efficiency as a function of block size and laid down a challenge to the theorists to explain it. There was a second one on Cyclic Redundancy Code (CRC) Polynomial Selection For Embedded Networks. This paper analyzed ALL 3-16 bit CRCs up to 2048 bit message size to find which ones were good in which regions. If you are selecting a CRC polynomial for a short block size, it seems that this paper is for you.

In my session, there were two talks in addition to mine. “The Recursive NanoBox Processor Grid: A Reliable System Architecture for Unreliable Nanotechnology Devices” presented studies of effectiveness fault-tolerance in various models. The second “The Impact of Technology Scaling on Lifetime Reliability”, developed phenomenological models for scaling of various macroscopic parameters of a processor based on feature size. They basically say that in the next couple of years, processors are going to get as hot as fire.

Finally, there were several wireless ad-hoc network papers. One I was particularly interested in “Evaluating the Impact of Limited Resource on the Performance of Flooding in Wireless Sensor Networks” (WSN group at UWA). They have a nice simulator for wireless ad-hoc simulators called Boris.

That about sums it up. There are a lot of nice talks, but these were the ones that caught my interest the most.