Jed Daniels . com|net|org

I work for a company called Network Physics. They make a pretty cool product, called NetSensory. Basically, you plug it into your network and feed some traffic to it (with a monitor/SPAN port, a network Tap, or a hub), and it calculates buckets of metrics about the data. Then you fire up the NetSensory Console (a java app that you download form the appliance), it sucks down the data and metrics for the time period you want, and lets you display them in about a billion different ways.

There are a lot of different products out there that I’ve played with for monitoring purposes, but NetSensory is really amazing with the variety of things you can do with it. You can use it to do all the basic stuff you are probably used to, like see who is using your bandwidth and what protocols are on the network, etc.; but you can also use it for so much more.

• You can find and debug application performance problems. Let’s say I’ve got some database servers that are feeding my website. If they start acting up and responding slowly, I can tell with NetSensory. I can also use it to tell if the webservers are appropriately spreading the load across the database servers. Or if the webservers are responding slowly or traffic is unbalanced to them
• You can manage changes in the network. Let’s say your boss wants to spend 50 grand on some load balancers: you can get the data necessary to determine if you should really do it, or if you should just cram a few a few extra sticks of RAM into your servers and pocket the remaining 49K (or I suppose you could let the company keep the money too).
• You can monitor your WAN traffic for usage and SLAs.
• You can spot worms and port scanners easily and track them down quickly.
• You can find and identify packet loss and other problems in the network.
• You can find rogue users or applications.
• And much more.

But, there is a problem: you’ve gotta know what you are doing. You can’t just launch the thing and expect it to tell you everything about everything, you need ot know your network and you have to actually LOOK at the data.

And, up until very recently, there was another problem: you couldn’t try it without convincing a sales critter that you are serious and then getting a demo appliance to put in your network (of which there are a limited quantity of, so not everyone could get one even if they were serious). But now, that isn’t so much of a problem, because we’ve released a version of the product that runs in VMware. The VMware version fills two rolls: it functions as a trial that ANYONE can download and use with minimal hassle, and it can be permanently licensed for use in a small office environment. The trial version only collects data for 5 days, but you can look at the data indefinitely, and because it is VMware, it is easy to reset and start another 5-day trial. It is also limited to only 5000 packets per second, so it will only work in small office environments (but larger environments can still use it as a test on a limited network segment, and can upgrade to a “full” appliance if desired).

So head on over to the download page and check it out. I think that most Network admins will be able to get some real use out of even just the 5-day trial, even if you never buy it. (Which sounds like something the company would be scared of, but they realize that the more people use the trial and do useful things with it, the more they will realize that NetSensory really rocks, and will eventually buy the product.)

So if you want to give it a try, head on over to the download page and give it a whirl. I think you’ll be happy if you do.

Some extra details you might want to know:

• The purchase of a license comes with Instructor Led Training (which you’ve got a decent chance of being taught by me).
• It also comes with a year of support. Yep, support is included.
• And perhaps one of the coolest things: NetSensory has these things called Insights that are used to display customized data in specific ways. YOU can write your own Insights that are custom for your environment, or download others that community members have written and shared from www.itsnotthenetwork.com/community-insights. These things are amazingly powerful, you can do a ton of stuff with them, including integrate scripts (perl, shell, whatever you want) into them.

NetSensory Virtual Appliance Trial Download from Network Physics

• Comments (0)

This weekend I taught my mother how to use the text messaging on her phone. She’s pretty smart, but like many people that haven’t grown up inundated with ever-changing gadgetry, she just doesn’t have the necessary background to “figure out” all this new fangled technology. She knows perfectly well how to type words into the phone, she does it all the time with her address book. For example, if she wants to create a new address book entry and label it “Richard”, she goes to her address book and types

777 444 222 44 2 777 3

(the spaces are there for your readability, she doesn’t actually type them). This spells out R-I-C-H-A-R-D on the screen of the phone.

But when typing a text message, it doesn’t work like this. If she tried to type the same word, “Richard”, she would enter:

777 444 222 44 (beep) 2 (beep) 77 (beep) (beep) 7 (beep) 3 (beep)

and the screen showed Sprighabag?. What? Sprighabag?? What the heck is Sprighabag?? Stupid technology. (Actually, after the first beep, she probably got a little worried and may not have even continued typing. If she did, she certainly didn’t get past the two consecutive beeps, and likely put the phone down convinced that it was out to get her.)

The problem is the phone is trying to be helpful, and it ends up outsmarting itself (or her, I’m still not sure which). When texting, it goes into a special mode that uses “predictive text technology”, where it tries to make things simpler by lessening the number of key presses necessary to spell a word (sounds fancy, huh?). Instead of behaving the way that it used to, it now expects only one key-press per letter, and it guesses the word intended based on the fact that there are only a certain number of possibilities for each sequence.

Lets look a very simple example, the word “hi”:

The “old” way:

44 (wait till cursor starts blinking again) 444

The “new” way, with predictive text technology:

44

That’s it. Since the number 4 only has three possible letters: G, H, and I; there aren’t a lot of possible combinations that are actually words. Here are all the possibilities: GG, GH, GI, HG, HG, HI, IG, IH, and II. The phone has been programmed to know that none of those other choices make sense, so it assumes you want the word “HI”. Easy, huh?

Let’s play around with it a little more, this time with a more complicated message. Get your phone and type this into a new text message (on my mother’s phone, a Nokia 6061, you get to the new text message screen by pressing up on the four way navigation button–your phone might be different). Seriously, go do this.

(Don’t type the spaces, they are just there for readability here):

8447 0 47 0 2 0 8378 1

What did that spell? Neat, huh? Go ahead, take a minute to type some other stuff. Notice that pressing 1 gives you a period? I’ll talk more about that a little later. Fun now that you understand how it works, isn’t it?

Now let’s look at a slightly more complicated example. What if you want to type the word chef? Type the numbers 2433, which would spell chef, but also spells other valid words, like aged and aide. The phone first gives you Aged, but you can press * to have it display another possibility. In this case you need to press * twice before it realized that you want the word chef.

2433 * *
gives you:
Aged -> Aide -> Chef

Remember how earlier we learned that the 1 key will enter a period? Well, you can use it in conjunction with the * key to get other punctuation as well. Pressing it will cycle you through the -, ?, and more.

Now what if you want a word or abbreviation (or name) that it simply doesn’t recognize? For example, you boss sends you a text with the question “When can I expect to receive those TPS reports? Don’t forget to use the new cover sheet!”

You want to respond with “I’ll have them for you eod” (Eod is an abbreviation for “End of Day”).

So you type:

455 0 4283 0 8436 0 367 0 968 0 363

But 363 gives you end, which you don’t want, cause you just want the abbreviation “eod”, not to type the whole thing. So you press * to cycle through the possibilities.

363 * * * * * * * *
which cycles through the possibilities:
end -> foe -> doe -> ene -> eme -> enf -> dod -> fod -> end

Uh-oh, no eod. Now what? Well, on most phones at this point there is an option on the screen for “Spell” (in my mother’s case it is activated by pressing the middle key in the navigation pad, but your phone might be different). So you press “Spell”, and it brings you to a single text box that behaves the “old” way: you press each number several times to specify a particular letter. So now you can type

33 666 3

and then hit “Save”. Viola! eod

What if you just want to go back to the old way for all of your text messages? Or what if you want to type a word in all CAPS? Can you do that? Yes. By pressing the pound key (#, aka hash mark, aka number sign, aka cross hatch, aka octothorpe, aka comment symbol, etc) you can change the typing mode.

There are several possibilities, which you can scroll through by pressing the pound key repeatedly, and identify by looking for the little [image] symbol on your phone (it may be at the top or bottom of the screen, depends on the phone):

Predictive Text, Capitalize the first letter of each new sentence (the phone knows that each word after a period or question mark is the beginning of a new sentence)

Predictive Text, no capitalization

Predictive Text, All CAPS

No predictive text, Capitalize the first letter of each new sentence

No predictive text, no capitalization

No predictive text, all CAPS

Number mode (allows only numbers to be entered)

There are a few more tips to cover, but you should have the hang of it by now, and the “advanced” stuff changes a bit more depending on the type of phone you have. I’ll write more about them at another time.

Happy Texting!

• Comments (0)

This week the Slashdot Poll is about least favorite cliches, and one of the missing poll options that we aren’t supposed to complain about is “This is the year of Linux desktop.” Even if it isn’t, there are certainly more reasons to start experimenting with Linux than ever, and there are more new people to Linux than there ever have been. While there are lots of resources out there for learning some of the basics and getting started, sometimes you just need a reminder while actually doing stuff about what is what. Codejacked has written yet another great article that lists a few commands that can help remind you. This is an article about commands that give you information about commands, and it is an excellent resource for newbies and experienced users alike.

Learning Linux Commands [Codejacked]

This article is copyright OPNET Technologies, Inc., and is reprinted from the original at www.itsnotthenetwork.com with permission.

• Comments (0)

[Update July, 2007: I know, I know, I've been terribly delinquent about posting the details from the show. I will soon, I promise. There will be two more parts: one with pictures and notes on construction, and one with lessons learned and other info. Thanks for being patient, check back soon for the next installments of the faraday cage series.]

This year at the Interop Labs we are building a walk-in Faraday cage as part of the VoIP over Wi-Fi demonstration. Since there are usually anywhere from 250-400 wireless access points “viewable” from the show floor, we need some way to remove all the junk in the RF spectrum in order to provide a controlled environment to test in and to effectively demonstrate the differences between VoIP Wi-Fi connections with and without QoS controls and burdened and unburdened Wi-Fi networks.

Anyway, since this walk-in size Faraday cage project was my idea, somehow I got saddled with making it happen. So the first thing to do was to build a proof-of-concept before ordering the necessary materials. On the last day of hotstage (about a month before the show we get together for just over a week to do the planning and build out of the Interop Labs booth, called hotstage) I went to the hardware store, got some aluminum mesh (the kind you put on a screen door or window), some 1/4 inch hardware cloth (something you might build a rabbit cage out of) and folded them up into a couple of box shaped containers to put something in that could measure radio waves. Then I fired up Kismet on my laptop and put it in each one. It was a spectacular success. Moments after placing my laptop inside the aluminum screen “box”, the signal went away. Unfortunately, after running several successive tests, it appeared that the signal went away about 90 seconds after launching Kismet, no matter what the circumstances (I could be standing next to the access point and it didn’t seem to make a difference.

So I went on the hunt to find some other Wi-Fi measurement tools to run on my Macbook Pro. Of course the obvious choice is Macstumbler, but it hasn’t been updated in a long time and crashes at launch. Then I found iStumbler and APGrapher, and both of those seemed to work well. So I fired them up, put my laptop into the proof-of-concept device, and…the signal didn’t change. Buh.

Ok, back to the drawing board. A little more research and a refresher in high-school physics reminded me that a faraday cage is basically a big “ground” wire extended into a 3D space that you could put stuff in. Doh! I didn’t ground any of my proof-of-concept tests (well, actually, I did hold it up against a metal wall socket at the warehouse that should have been grounded, but that didn’t seem to make a bit of difference). So I set out to build a new proof-of-concept, and this time I would be more meticulous about its construction and I would make sure it was well grounded.

I decided that this time, I’d build a frame, attach the mesh to it, and secure a copper wire to the mesh that I would then connect to the ground plug (third hole) of a standard wall socket. I choose to build a box roughly 18 inches X 12 inches X 12 inches, so I could put my laptop inside of it without much trouble.

Step one: buy the necessary components. Here is the list (pics down below in the assembly instructions):

1. Two 8′ lengths of 2″ X 2″ wood (I choose spruce “furring” strips, cause they were $1.99 at the local hardware store). This will be the main frame.
2. One 8′ length of 2″ X 1″ wood (again, spruce furring strips, even cheaper at $1.79). This will be frame for the hinged lid.
3. Hinge to attach lid to frame. I choose a 12″ piano hinge.
4. Aluminum screen to cover. I made some rough calculations and bought 10′ of 36″ wide mesh.
5. Copper wire for the group connection. I choose green 14 AWG stranded, cause it was reasonably cheap.
6. A bolt, some wing nuts, and washers to secure the copper wire to the mesh.
7. A plug the I could attach the copper wire to and insert into a socket.
8. Screws and staples to secure the frame together and the mesh to the wood (I simply used what I had laying around in my garage).

Step two: Cut the wood. I cut the 2″x2″ furring strips into four 18″ lengths and eight 12″ lengths. And the 1″x2″ strip into two 18″ lengths and two 12″ lengths.

Step three: Screw the frame together. This isn’t rocket science, so I’m not gonna give you detailed instructions on this part. Just remember to drill pilot holes for your screws so you don’t split the wood. Here is the assembled frame:

Step four: Build and attach the lid. I simply screwed the 1″x2″ pieces I had into a square and put them on the frame box with a piano hinge:

Step five: Attach the screen mesh. I will never regret the day I bought my air compressor and pneumatic stapler. I can only imagine the terrible hand cramps had I attempted this with a standard spring loaded stapler. If you don’t have a pneumatic stapler, I highly recommend you at least look into purchasing an electric one. To make sure that everything was tight, I tried to keep staples no more than 2 inches apart (usually about 1 inch, but towards the end I got a little tired and lazy, so the space between increased a littler). Here is the before and after:

Phoenix the cat seems curious about the frame and screen.

Step six: Attach the ground wire. Below is a shot of the parts mentioned above, along with Phoenix, one of our cats who was very interested in this project. Attaching the parts was pretty simple.

You will notice that I cut off the two prongs that get electricity. I wanted it to be perfectly clear that this plug is non-functional, and won’t create an electrified cage (cause that is a completely different project!). [NOTE: This is dangerous! Don't do it. If there are some cables crossed or your ground is wired incorrectly, you could create a lethal situation and die. Get an electrician to get you a dedicated ground wire to use for a project like this. If you kill or seriously injure yourself, don't blame me, I told you not to do this. Seriously, I'm not joking. Don't plug stuff into an outlet that connects to bare metal that you will be touching. That is just stupid.]

Step seven: Time to test! I plugged in the ground, got my laptop and started measuring Wi-Fi signals. The pictures speak for themselves:

DOH! It still has signal. WTF?!? Well, as it turns out, the ground connection in my house is quite bad. So I brought it into the office and tried it there:

Woohoo! Just to make sure though, I opened the door, let it reacquire signal, then closed the door again:

Success!!!

Check back after the show (May 20-25) for Part 2, where I will detail building the walk in cage.

• Comments (3)