Obfuscated Networking is just a place to store ideas and thoughts. It's not clear what, why or how, but they are all here. Hopefully you will find them entertaining, thought provoking or interesting and will be inspired to leave your thoughts.

I’m digging through a lot of statistics on self similar processes to understand network response time variance. It appears that response time is self similar, that is, the graph of the distribution looks similar regardless of scale, which presumes that response time is directly tied to traffic levels in terms of PPS which is in turn tied to the distribution of file sizes and how often requests are made, which are both self similar (though one article claims there’s no correlation between the rate of requests but I’m not sure it applies in my situation). In any case I’ve added another factor in my network, the state-full firewall. This appears to change the distribution for a given quantity of traffic such that the more states a packet needs to be compared with the larger the variance in response time. I’m not sure, however, about the average response time however from the data I have so far that doesn’t seem to be likely, and if it is it’s not nearly as much as the change in variance.

What does this mean? If my data holds up (anyone know how to calculate the margin of error for the difference between two standard deviations for dissimilar sample sizes?) it means that you can reduce variance in your network response time by increasing the aggressiveness with which you timeout connections in a state-full firewall. This is particularly interesting if the average response time does not change.

One problem I’m working with right now is that one of my firewalls seems to have two periodic hiccups which result in 19.7 and 57.8ms response times from a locally connected host at some interval regardless of other changes. That is, if you graphed the distribution of the response times you would see two “bumps” on the graph at those points. I’m not sure how to explain this but since those values appeared in my control they should be accounted for in my new samples. Despite this the results still seem to show lower standard deviations of response time.

If I can correlate total search distance, that is the number of searches time the length of the table to changes in the standard deviation and I can correlate the traffic quantity and the distribution to figure how long of a delay I can expect to see x% of the time I can estimate when I need to add a firewall to my cluster of firewalls to maintain a particular service level.

Here’s the kicker… …I think I know what I’m talking about but I know just enough statistics to be dangerous, oh so dangerous.

In Rousseau’s Confessions the author attempts to convince the reader that, among other things, the horrors of his life are greater than they are; to this end he uses emotional imagery to create context for the reader in which the facts of his life will be interpreted subjectively in an emotionally supportive light rather than the cold, harsh, and objective light of reality. In Part 1 of Confessions the narrator speaks of his parent’s life before his own, explaining the love his parents have for each other and of their greatness; however, if you look closely at the ordering of events you can see a subtle manipulation of the reader. While each piece of information read in his order may be considered naiveté the careful structuring of the passage gives rise to a suspicion beyond doubt.
(more…)

So I have a couple of machines with 4 GB of ram or more running a 32 bit version of windows. According to Google and everyone everyone else I read I’m never going to use the full amount of RAM and there’s nothing that can be done about it with a 32 bit OS.

I’m calling Shenanigans.
(more…)

Amazon.com likes to bundle things together and offer you a bit of a discount. You’ve seen it, “Buy this book with XYZ other book now and Save!”

Today they’ve just gone off the deep end with me. The Basic Practice of Statistics 4th Edition is available with The Basic Practice of Statistics 3rd Edition together! Just what I’ve always wanted. The best part? No discount! This must be a match made in electrons.

Has anyone noticed how bad Amazon.com’s layout is? This is the most poorly designed website I use on a regular basis. There’s entirely too much information in a small amount of space and finding the piece of information you want is almost impossible. Let’s see how this works.

When I’m looking at an item there are two rows of tabs across the top of the screen, though only half of one looks like tabs. The first appears to be global site navigation options and the second appears to be category options within the selected tab. Directly beneath that is another row of options involving two search options, Amazon and A9 Web Search, and an option to buy Gift Certificates thrown in for good measure.

Below that is the main corpus of the item I’m looking at in three columns. The first is a picture of the item with options to “See larger image”, “share your own customer images”, and (since I’m looking at a book) an option for the Publisher to learn how customers can search inside this book. I have to wonder how many times a Publisher of a book looks at the book on Amazon.

The second column tells me some useful information such as the name & author and how many stars the book has from reviewers with an option to read the review. Below that is the List Price, the actual Price, how much I save, availability, options for having it delivered soon and a link to other sellers who have it New & Used.

The column to the right allows me to buy the book from Amazon, Buy it from other sellers (again), allows me to sell it myself and has buttons for adding it to a wish list, a shopping list, a wedding registry a baby registry and telling a friend. I’ll be sure to add Statistics, 10th Edition to my baby registry when I’m having a child.

Below all this there’s a Product Promotion letting me know that I can get free two-day shipping on this item when I spend $200 on qualifying textbooks ordered by Amazon.com. Below that is an opportunity to buy it with another book to save money.
(more…)

Via the cool people at Gizmodo I found this. No longer are model airplanes just those dorky propeller planes. Now they are turbine driven military fighter aircraft. This is a 1/5 scale of a Soviet Mig. It just about needs a camera and video feed to be able to pilot it because it flies so fast. Watch it’s takeoff and see how fast it accelerates down the runway.

Link

Some builds of Apache 2.2.6 on FreeBSD 6.2 may fail when ld can not find libexpat. One solution is to use the built in expat however I consider this suboptimal. Fortunately I remembered my historical joys of building software from the days before package management and ports. I’m going to presume c-shell because that’s what root uses. Translate as necessary

# setenv LD_LIBRARY_PATH /lib:/usr/lib:/usr/local/lib
# ./configure <YOUR OPTIONS HERE>
# make

All should run smoothly presuming you’ve installed the Expat port (/usr/ports/textproc/expat2) or have it installed via some other method. Happy building.

Apple has a Broadband Tuner application that can be downloaded to increase performance if you are on a very high speed network or on a network with high delay (such as a satellite link). In both these situations the amount of data that might be on the network between the two computers is very high. Because of the design of TCP this requires larger buffers. The combination of bandwidth and delay is called the Bandwidth Delay Product or BDP.

The Broadband Tuner page incorrectly identifies FiOS802.11B networks might only have between 1 Mbit/sec and 5 Mbit/sec of bandwidth in real world usage however 802.11N draft products like Airport Extreme have provided over 100Mbit/sec of bandwidth which is more than enough to take advantage of FiOS’s 20Mbit/sec (or more).

From the good folks at Cruel.com comes The Truth About Vaccines. I don’t know whether to laugh or cry, but fess up. Who’s been practicing Witchcraft around here? I want in!

Also remember, King James only. Accept no substitute. Note the link to human sacrifice.

On FreeBSD the sysctl kern.ipc.nmbclusters is used to limit a particular system resource. If you set the value to 0 in /boot/loader.conf the resource is supposed to be “unlimited” in recent versions (at least since 5.3). I seem to have found a strange bug. If you set this and you attempt to mount an NFS file system using UDP you can mount the file system but read operations fail. The request is sent and the reply is received by the OS (as seen in tcpdump) but the NFS client code does not return the data up to the requesting process.

I’m currently watching the system to see if other issues associated with this server appear.

“One man one vote.” is heard whenever people try to institute voter ID laws or other measures that make it more difficult for people to vote. Let’s see how difficult it is to vote twice in the Texas Legislature.

I imagine that much of this is done with agreement between people so that they are being voted for in the way they intend but I wonder if anyone’s really paying attention to the results. I guess in Texas the phrase, “I don’t recall that.” is actually a true and valid answer since apparently many didn’t actually vote as the record shows they did.

Maybe we should require biometric identity scanning or put a RFID chip in our legislators so they can’t vote twice.

Microsoft Excel 2007 has a slight problem. Click on the image to see the enlarged photo. If you perform some calculations where the answer should be 65535 you get a result that’s clearly not 65535. I don’t know how this one got past QA. I don’t even know how they intentionally got this one wrong. It’s not a processor bug like some previous problems were. This one just blows my mind.

This appears to work for other values such as those same values but one divided by 2 and the other multipled by 2 such as 38.55 * 1700. I iterated that several times in each direction. I attempted to do it further by using Control + Down Arrow and it maintained the same result until I got past 1*10^308, which seems to be some limit in Excel.

Now, I put the value 65535 in cell A1, =(A1/850) in B1, =(A/1B1) in C1 and =(B1*C1) in D1 which should perform the same multiplication but it returned 65535, which tells me Microsoft is smarter than your average bear when it comes to calculating cells and is sure to reduce values when it can.

4369/3 * 45 yields the correct result. The first part is a repeating fraction.
225*257 yields the correct result.
128.5*510 yields the correct result
51.4*1275 yields the correct result.
So far the only numbers that give an incorrect result are related to 850*2^n or 77.1*2^n where n is a positive integer.

850*77.2 and 77.1*851 both yield correct results, so this tells me it’s not related to those two numbers directly.

Update: Apparently Someone who used to work on developing Excel has insight in to the problem. Apparently the math with those values still works properly, it’s only the display of those values which is incorrect which you can verify by making one cell = to 77.1*850 and another cell equal to that resulting value divided by 2. You get 32767.5.

There is also a MSDN Blog article about this.

So you’ve got a well used FreeBSD machine and you want to build a system that is similar in build. The question becomes how to handle ports. If your system is anything like mine you have a ton of ports installed. You can login to both systems and list packages on one and find then install the packages on the other. After the first four this gets annoying. Enter your pkg db.
(more…)

I like virtualization technology. It lets you do bad bad things. In this case I’m running an iMac running Parallels running Windows XP running VMware running FreeBSD running X11. I hope to get Linux compatibility running and start VMware for Linux on FreeBSD so I can run Solaris X86. click on the image to see more screen shots. I’ll add more over time.

While discussing the translation of poetry on a mailing list usually devoted to guns, politics and religion a former associate and current good friend of mine, Andrew Lagemann, wrote an interesting essay. First some backdrop

The original writer wrote

But one can translate Japanese poetry very effectively.

Poetry that focuses on capturing a moment, as Japanese poetry often
does, is often effectively translated.

Andrew responded with the following

(more…)

Every once in a while a translator will attempt to translate a piece from on language to another with the utmost care to convey the meaning to a different culture. This is particularly difficult when translating between languages that don’t share common written, spoken and grammatical structures such as Japanese and English. One such attempt is the translation of Matsu Basho’s “Narrow Road to the Deep North” by Nobuyuki Yuasa. This attempt is the clearly the most insulting translation of Japanese literature I have come across yet, particularly the translation of the poems.
The problem with this translation is obvious from the first poem. The poems were originally written in a three line form which predates the use of the term haiku but appears to be a predecessor of the form. The translation in our text reads

Even my grass-thatched hut
will have new occupants now;
a display of dolls

While this likely looses some aspect of the original it retains the sense of poeticism imbued within and without. While it’s not always possible to translate the varied meanings of words used poetically this translation is, at least, poetic. Compare it with the following from Yuasa.

Behind this door
Now buried in deep grass
A different generation will celebrate
The Festival of Dolls
(more…)

So you install Parallels 3 on your Mac, you install Windows XP, you run dxdiag and things look neat, but no go with the game; you receive an error telling you that you don’t have something such as Pixel-Shader 1.1 despite having a < DirectX 10 video card… Stop your instance. Go to Edit->Virtual Machine to open the configuration editor. Select your video card and enable DirectX support. Blam-o, things work.

The Shakespeare sonnet “Shall I Compare Thee To A Summer’s Day” contrasts the subject of the poem to a summer’s day. While many aspects of a summer’s day may make for popular poetic fodder this day does not fare as fairly as in other works. The poem starts with the narrator asking the lead the lead-in question “Shall I compare thee to a summer’s day?” and follows with the first comparison “Thou art more lovely and more temperate.” He then describes summer in following six lines as a short, sometimes rough and at times overbearing event whose fairness and beauty is too often overshadowed and which ends all too quickly. This is followed by a contrasting description of the subject in the following quatrain. The narrator ends with the couplet “So long as men can breathe or eyes can see / So long lives this, and this gives life to thee.” This is the only information available to give evidence to the subject and the comparator implying that the subject is something that cannot be openly displayed and thus requires the words of the poem to have life. Short of the ending, the comparison gives scant idea as to what the subject is. Shakespeare’s sonnet “Shall I Compare Thee To A Summer’s Day” is an abstract poem where the narrator could be saying almost anything about almost everything .
(more…)

Having a magic number in August has been nice. Having the best record in your league by far is nice. Being tied for fewest losses in the majors is nice. Somehow though people can’t help but to diss the Mets. The big focus has been on their pitching. No good starts they say. Ace reliever out for the season they say. Closer is shaky they say. No one pays attention to the fact that the team with the most runs in the National League is also the team with the best ERA in the National League. Pitching? Who’s got better pitching in the National League? I’ll give you that Detroit has better pitching, but Detroit can’t hit the way the Mets can.

Sure the rotation has issues, but it’s had issues all season, and what happens? Everyone else steps up to the plate (or is that steps up to the rubber?)

You can’t deny the fact that this team doesn’t give up runs and wins games. For whatever perceived weakness you see in the pitching staff the other 8 guys on the field more than make up for it both with their bats and their fielding.

Oh, and when the World Series does come around the Mets have a legitimate DH, Julio Franco. Lets see the AL’s pitchers hit as well in NY as he will in their parks.

Oh wait, I hope none of you nay-sayers meant to say someone like St. Louis or LA would beat them? What’s the record against the division leaders? 6-3 I believe with three games at home against LA coming up.

I just realized what my cars name is. It’s Precious. It’s my Precious….