top image
home  /  pages  /  tech tips  /  contact about

Paper, plastic, people and problems

March 25, 2000

Hi all,

I am writing this from my latest acquisition: a computer. MIT gave me a computer to use at home! I'm really happy. Now I can work 20 hours in stead of only 14. It was amazing: a day after I mentioned it I had a running system on my desk in my bedroom (as you can clearly observe, I have lowered my standards significantly as to what makes me happy in my bedroom). The hardest part of the whole undertaking was to get the machine physically out of the building past the guards. It is kind of suspicious walking out the building with a computer under your arm. Now I feel like a normal human again: my small laptop can now rest in peace while this newly baked silicon is humming cozily working the living daylights out of itself.

My research is going so so la la. My biggest problem is a serious lack of feedback. I'm basically all on my own. I knew nobody was going to hold my hand or operate the computer for me, but I feel a little like a blind guy in the middle of Manhattan. I have absolutely no clue what it feels like to be blind in Manhattan, but I'm sure it's not a very worthwhile experience. Nobody told me my idea was a good thesis subject nor did anyone say it was bad. As long as nobody stops me, I'll do it myyyyy waaaayyyy.

As you know I am trying to come up with (a) solution(s) for Denial of Service (DoS) attacks. Now you have to know that whatever solution I will think of, it will be public (since all the software we write is public). The bad guys (women don't do no naughty shit) will know as much as I do and they will surely try to find a way to circumvent whatever I come up with. Knowing this, I have to split my mind in two: one half tries to think of a way to stop DoS attacks, and the other immediately tries to find a reason why that won't work. (Kind of makes me feel like the main character in die Schachnovelle from Stefan Zweig. He got locked up in a cell and started playing chess against himself in his mind. Drove him totally nuts.) Usually there are at least 3 reasons why each plan won't work. The best thing so far seems to be a cocktail of all my ideas, but still it will only slow down the attacker, not stop him. Sadly, it's much easier to be the bad guy than the good guy.

In simple terms I try to define what a router can consider to be 'normal' traffic passing by. (Remember that a router is a computer that gets packets on its inputs and sends them away to one of its outputs (inputs and outputs simply being cables), the output depending on the destination address of a packet). I have come up with a scheme to enable a router to collect certain statistics about the traffic it is handling without taking too much time and resources, since you don't want routers to spend too much time on other things than routing since that will slow down the traffic. Based on these statistics I try to define how an attack will differ from 'normal' traffic based on predictions of what a possible attack will look like. To do this I had to read a lot of material about attacks that have been mounted so far and try to understand what all (or most) of them had in common. (I have some fun ideas for new Denial-of-Service attacks, by the way. Really inspiring reading material.) This mechanism could enable a router to determine whether certain traffic is (part of) an attack or not. After analyzing data (60GB, remember the phone books) I found out that there is, most of the time, indeed such a thing as 'normal' behavior, but unfortunately there are not-so-rare exceptions that make my whole idea shake on its foundations. I'm slowly accepting the fact that this idea will probably be only part of a larger solution, but not the remedy itself.

There is a significant difference between making friends with Europeans and Americans. In Europe, the hard part is actually getting to talk to one another. Once you're on speaking terms it is quite easy to move on to friendship. This is totally different here: making contact is no problem at all. Nobody considers it to be strange to just start babbling to a perfect stranger. The becoming-friends part is another story, though: apparently it takes months. It took me a while (pronounce 'some disappointments') to discover that.

As I was clipping my nails a few days ago, Martin came up to my closed door and we had this - what I found to be very amusing - conversation:
M: "Thomer?"
T: "Yeeees?"
M: "Are you clipping your nails?"
T: "Yes, Martin. Would you like to have them when I'm done?"
M: "No, it's just that I now understand the Dilbert cartoon about the 10 most annoying sounds coming from the cubicle next to your own."

It took me 2 minutes to stop laughing. You have to know that the wall between Martins room and mine is really thin. If I pick up the phone and it's for Martin, I just talk to him as if he was standing right next to me. He always answers. Quite similar to the God concept, actually.

My house is rather small, you know. This is usually no problem since most of us are not at home during the day (Dan the Purring Hairball is here all day crashing his head into furniture. Sometimes I hear a soft *THUMP* somewhere in the house: it's Dan head-banging again) so weekdays are no big deal. Problems arise in the weekend, though. Problems is maybe the wrong word for what really should be called a genuine in-house traffic jam around bottlenecks such as the refrigerator or the washing machine. Oh yes, speaking of traffic jams. (Time for some complaining; makes my Dutch roots blossom.) I am absolutely stunned by the lack of speed with which cashiers do their work here. In my Star Market at Porter Square in Somerville you can pay at 15 different checkouts, but they - as a collective - are as slow as 2 checkouts in the Albert Heijn. If you know that at every checkout there is one person moving the products over the laser and someone else putting your stuff in plastic/paper bags (about which I will elaborate extensively in the next paragraph) it is absolutely incomprehensible how they manage to stack up queues of more than 5 people per checkout. The worst is that nobody seems to mind, which I find baffling for a society where time equals money. Apparently it is a well established fact that you just have to wait at least 10 minutes to get to the front of the queue.

As promised: the Bag Situation in the United States of America. People transport their stuff from the grocery store to their home in small plastic bags they get at the store. The basic concept is that the plastic bags are too thin to cope with heavy objects such as 2 liter Coke bottles (minimum size, no kidding). Solution: put a paper big IN the plastic bag to prevent the bottom of the plastic bag from ripping open. This solution has one glaring flaw, though: the handles of the plastic bag can't handle the weight (makes you wonder about the appropriateness of the word.) The handles break when there's too much in the bag. There always *is* too much in the bag since there is always too much in everything (and everyone) around here. I discovered this the hard way when I lost one of the bags from my bike while speeding down a hill at a velocity of 30 km/h. The bag let go of my steering wheel. Murphy made sure that the only glass pot I bought that day was in that bag. Of course there is a simple solution to this exceedingly complicated problem: put the whole paper-bag-in-plastic-bag construction in another plastic bag to divide the weight over 4 handles as opposed to only 2. To paint the whole picture: a small set of articles is put in a paper bag, and the paper bag is put in (at least) two plastic bags. Please bear with me now while I elaborate a little on this by means of another calculation, this time not involving telephone books. Say that you can fit 5 average sized articles in a bag-composition like that (this number is what I experienced to be the average, not considering the fact that they wrap eggs, meat, ice-cream and some other articles in a separate plastic bag before putting it in the p-b-i-p-b construction). Let's say the average person goes to the Star Market 2 times per week and buys 25 articles each time. That's 10 paper bags and (at least!) 20 plastic bags per week per person. There are 250 million Americans. Let's say (considering the number of articles per week) this is for a family/group/set of 3 people. Hang in there for just a little longer. I guess 50 articles per week for 3 people is a reasonable guess. That's 830 million paper bags and 1660 million plastic bags per week! Why does that make me feel all creepy inside? We (at home) don't throw the bags away because of the abundant space it takes in the garbage, so we collect the bags. It is scary to see how quickly our collection is growing. Extrapolating our private collection to the whole nation makes you wonder.

As for the contest I gave you last week (all the abbreviations): two people took a shot. Xander did the best job so far. He got 23 right and 2 wrong. My brother-in-law got about 20 right. The contest closes at March 31, so if you want to apply for the red-lipstick kiss postcard or even for the curl: don't wait, call now. (1-800-GET-CURL, getcurl.com).

Enough nonsense for now. Bye,

Thomer

P.S. Some of you have expressed worries about the time I spend writing these e-mails. To give you a rough idea: this mail took me 4,5 hours to write. Let that be proof for how much I care for you all. Americans would simply call me anal. Anyway, add another 30 minutes tomorrow to edit it to make it suitable for the Web page. I started writing this one at 12 at night... It's very late so I'm going to bed now.

URL: http://thomer.com/mit/9.html
Copyright © 1994-2011 by Thomer M. Gil
Updated: 2004/09/06