1. Windstream’s biggest customers are AT&T and Verizon
Telcos buy backhaul from other telcos when it comes to providing cellular service. As long as Verizon and AT&T continue to rule the roost in the wireless biz, Windstream will continue to get a fair chunk of change per month for T1s and fiber Ethernet connections to cell sites in their footprint. In many cases other cellular providers also have to use Windstream for backhauling their cell sites, which generates even more revenue for Windstream, which isn’t particularly forgiving when it comes to T1 pricing either.

2. Rural wireline customers are going to be around forever
His words, not mine. The idea is that cellular service out in the sticks isn’t all that great to start with, and  if cell service is good then it’s riding over Windstream lines. Rural folks are also more reticent to give up landline service even if cellular service works just fine, and Windstream DSL  goes places that cable never will. Combine that with a USF kickback or two, plus low pricing on both DSL and landline service and you’ve got, in theory, a captive user base who will keep paying you $5o or $60 per month ad infinitum, as long as you keep that copper pair to their house in reasonable shape. Oh, and you’re getting some USF funding on each line because you’re rural; even Windstream’s “naked” DSL service still has a phone line involved, albeit one where outgoing calling of any kind costs 10¢ a minute.

Of course, the model breaks down if you’re serving areas that do have cable, where you just have to compete on price or lose customers unless the cableco is grossly incompetent. Or if the cellular providers use the cablecos for bandwidth and take wireless backhaul routes instead of Windstream wired ones, Windstream loses out. Or if some wireless ISP decides to offer service that beats Windstream’s standard 3 Mbps down, 635 kbps up (after ATM overhead), $40 DSL service, Windstream’s position gets a little less tenable.

Of course, Windstream does do fiber to the home for all of their new builds, and the tech supports either 12 or 24 Mbps down and 768k or 1.5M up. However even those speeds pale in comparison to what companies like Comcast have done with their DOCSIS 3 products. Then again, Windstream tends not to compete against Comcast.

3. Windstream could have bought Frontier, FairPoint , Embarq or even Qwest, but the price was too high
Windstream is no stranger to acquisitions; buying their way into being the ILEC in a given market  has been done several times recently, with Iowa Telecom being the biggest purchase. Windstream is also not averse to buying competitive carriers, as long as they can make a pretty penny on data revenues. To that end, they bought NuVox.

However the company has passed up deals like Verizon’s now-Frontier territories, the sinking carcass that is FairPoint and even such hulking giants as Embarq and Qwest, leaving the two last companies to CenturyTel…erm…CenturyLink. What’s interesting is that the rural telephone market…which also happens to include Las Vegas, Phoenix, Denver and all of Oregon…seems to be converging to two or three carriers: CenturyLink, Windstream and Frontier, with Frontier a possible acquisition target down the road by one of the other two operators. Note that the consolidation talk reflecting future buy-outs is my conjecture, not that of the source I talked with.

4. Verizon will sell its remaining rural markets, but they may be bled dry before that happens
Verizon didn’t sell their rural Texas markets to Frontier in the gigantic July 1st swap, yet many of these markets will never see FiOS. So Verizon is shopping the markets around…somewhat…though at this point the price to Windstream is still much too high for the company to bite. In all likelihood, Verizon will wait too long to sell to a provider who actually cares about copper-based service, forcing the company to offer fire-sale pricing because they are selling copper plant rather than a customer base. WIndstream is good at keeping churn low, but I don’t hear about the company picking up new lines of service on copper infrastructure unless it’s DSL service for someone who already has a landline.

So, what do y’all think about the above?

The problem in my case had to do with embedding videos into the slideshow. You see, when Picasa exports a slideshow it takes advantage of some of the more esoteric specs of the MPEG-4 container, saving video clips, unmodified, with their accompanying audio alongside the photo slideshow, which includes your music track, captions, etc. So you get multiple video and audio streams “neatly” packaged into a single MPEG-4 container. Cool, right?

Right…as long as you have something that won’t roll over and die when it sees a multi-stream, multi-resolution container file coming down the pike.

Yes, I said multi-resolution; the camera that I used to take the videos built into the slideshow shoots at 720p. THe slideshow itself was outputted at 640×480. Add to that multiple audio stream formats (the music track was one thing, the videos were another) and you get a rather fragile system if your player doesn’t support the whole MPEG-4 container spec to a tee.

QuickTime took the whole thing in stride, though of course deciding 720p video on a 1.6 GHz Pentium M laptop is not as easy as dishing out 640×480. However every other application, transcoder (e.g. ffmpeg) or player (e.g. VLC…yes, even VLC didn’t get the clip right) either couldn’t deal with the video at all (ffmpeg) or was missing one or more of the audio or video tracks (YouTube and Vimeo…neither service transcoded the clip correctly).

Yes, YouTube butchered a video that was exported from Google’s own photo management software. They need to work on fixing that. I thought the issue was with the copyrighted music I had playing in the background (MercyMe’s All of Creation), but they’re actually perfectly fine with that, as I would expect Google to be. After I uploaded the video in a format the YouTube transcoder could parse correctly.

What was that format, and how did I get the darned thing to transcode correctly? Wait for it…MPEG-4…from QuickTime, via iMovie.

I imported my slideshow output file into iMovie (the new version, ’09 I believe), then used the Export to QuickTime option to get some fine-grained control on exactly what I was doing. The specific settings I used were as follows, though minor changes shouldn’t mess anything up…

• Video compression type: MPEG-4
• Frame rate: current
• Key frames: automatic
• Data rate: restrict to 1500 kbps (for quicker uploads over my 384k up DSL connection in TX; quality is still fine)
• Compressor quality: best
• VIdeo filter and size: no changes
• Sound format: AAC
• Channels: Mono
• Rate: 22.050 kHz
• Quality (under advanced settings): Best
• Bit rate: 64 kbps
• Encoding Strategy: Average Bit Rate
• Prepare for Internet Streaming: unchecked

The above settings got me a video that uploaded perfectly to YouTube, though it wouldn’t work right on this site’s (brand new) FlowPlayer install. Can’t win ‘em all I guess, though I’m sure for people exporting their slideshows from Picasa to YouTube their service is plenty.

When will Google fix Picasa’s multi-res MPEG-4 export issue, or tweak YouTube to import the videos without making scrambled eggs of the lot? Not sure. Maybe a few dozen retweets and a Digg posting will do the job.

Comment if you’ve run into this issue as well, and let me know if this fix worked. Also, if Picasa on Windows has this issue, let me know; I can test the process on a Windows box but haven’t done so yet. Unfortunately though, if the same issue exists, I know of no Windows application that can transcode its way out of this situation. Unless Apple still does the whole QuickTime Pro thing for Windows, post-OS X 10.6.

The server also seems to be more snappy; not sure whether it’s because the Virpus server is packing more VPSes onto a machine stocked with four dual-core Pentium 4 based Xeon chips (at 3 GHz) versus the QuickWeb Xeon 3220 VPS count or what, but the QuickWeb server feels…well…quicker. Sure, I’ll be limited to 1TB of transfer (yeah I know, *limited*) per month (versus 1500GB for $7 or 3000GB for $12 at Virpus) but that 1TB can be transferred over a fire hose of a connection: a gigabit port, to be exact. Needless to say, if I need to deliver some files quickly to a bunch of folks they’re going on this server.

But back to the title of this post…here are some quick VPS-related tips…key words will be linked…eventually…

1. If you’re doing file serving, do it with nginx. It can be installed via the command line from apt-get and runs smoking-fast. I’ve heard that the Cherokee web server is similarly awesome but haven’t tried it out yet.
2. Need a “virtual desktop” experience for web browsing? Install xfce4 and tightvncserver packages. If you don’t like GMail freaking out on you (saying your browser isn’t supported) and/or want to use Flash without bending over backwards, grab Firefox. For something fast, light and still compatible with most sites you come across, grab Midori. For a text-based browser, Links is awesome. For Firefox, the flashplugin-nonfree package will get you Flash goodness. NOTE: Chromium (Google Chrome’s geeky twin), when I tried it, didn’t work.
3. If you want to change the text displayed when you log into your system via SSH, edit /etc/motd (make a backup first, just for kicks). You’ll need to restart your SSH daemon (/etc/init.d/ssh restart) to get the changes to register.
4. To change your server’s host name (for example to align it with the domain name you’re using to access the server) type in hostname NEWSERVERNAME (including the domain) and you’ll be good to go.
5. To change your root password, log in as root and type in passwd. Yeah I know, that one’s a gimme but never hurts to state the obvious, right?

Y’all have a nice night now, ya hear?

1. Symmetric encryption (AES etc.) is fast but you have to get the key from point A to point B securely
2. Asymmetric encryption (RSA etc.) negates the key sharing requirement but is slow
3. Asymmetrically encrypt a symmetric “session” key, then use the session key to encrypt the file/message you’re trying to send and you get the best of both worlds!

So I made a command line utility, using the Chilkat Visual C++ library set, that did the above for file encryption, with automatic signing built in so the recipient knows who sent the file. The application is rather basic, but it works very well (and pretty quickly for encryption/decryption…my IdeaPad took about seven seconds each to encrypt and decrypt a 100MB file) and is easy to dissect for anyone who wants to do that sort of thing.

For more information, download the package below, which includes the Visual C++ Redistributable (I think it’s required to run the encryption/decryption applications, but you might already have it), encryption and decryption binaries (EXEs, sorry non-Windows folks), sample public and private keys, the source code for the project (all ~120 lines of it) anda report about the application. It’s an EXE file because 7-Zip does a much better job of compressing the encryption/decryption binaries than plain old Zip, meaning that the file, self-extractor included, is still 500+ KB smaller than the equivalent ZIP file. In other fun news,both encrypt.exe and decrypt.exe are under 1400 KB each (floppy size!), both sipped are around 1100 KB, and both 7-zipped (without self-extractor) are less than 500 KB!

Questions and comments are welcome…post below. As far as licensing on this goes, just attribute me and tell me that you’ve done so. Thanks!

Download Package

I should also note that I ran for Student Body President here at Mines. This isn’t the first time I’ve run for such a position; I ran last year for Junior Class Treasurer (a friend was running for Junior Class President and I didn’t want to split the ticket), and in fall 2007 I ran for Freshman Class President. This is the first time I was soundly defeated; the other two positions were won by a handful of votes by my opponent.

To be clear, I’m not bitter, just solidified in the supposition that the whole thing’s just a popularity contest; people vote for their friends, not on issues. For what it’s worth, I’ll probably end up doing nearly as much this coming year through <non-presidential ASCSM position> by saying “yes” to tasks that the guy who won wants to delegate. Or that I think up.

Sorry for the broken links in the tweet; didn’t see any use keeping up the Facebook campaign page that I had up. If you want to see the campaign poster I had, comment (Openness, Efficiency, Free Bagels were my three legged stool of stumping prowess).

UPDATE 3: Added another news source. Swamped by homework so tweets will have to wait, but other than those I think I’ve got all the online info on the outage linked from here.

UPDATE 2: Added another couple of news sources, corrected one of my usual typos…some how I forgot to add a “not” into the paragraph about Wal-Mart taking credit card transactions. Sorry for the confusion. Also modified information based on an article from the Fredericksburg Standard Radio-Post. The link below takes you to the subscriber-only version of the article; the non-subscriber snippet just makes the paper look bad (very little info).

UPDATE: Added some more news articles, (still) working on collecting all tweets about the incident, revised some information for accuracy (originally wasn’t sure about CricKet and Pocket, but they too were taken down by the outage…DSL was also apparently iffy in Kerrville).

Granted, I wasn’t in the area when it happened, but since I have plenty of family and friends down there I’m certainly going to talk about it!

Kerrville Daily Times Articles
March 24th Front Page (JPEG image)
Photo
Editorial
Note about landlines/safety
Info about 911 operations
Main article (covers cellular, credit cards, banks) Reprint in iStockAnalyst
Businesses adapt (anyone remember credit card imprint machines?)
Humorous recap by one Daily Times reporter (not a newspaper article)

MySanAntonio
911 article
Original article

Articles from other outlets
KENS5 KVUE WOAI WOAI VIdeo KSAT12 Boerne Star Burnet Bulletin Fredericksburg Standard Radio-Post Blanco County News TMCNet

Rather scary stuff for the folks affected. The outage, by my calculations, lasted for a full seven hours, and knocked out cellular communications completely in the Hill Country. Verizon cells? Down. Sprint? Down. AT&T? Dead. T-Mobile? Not 100% sure, but probably gone. HCTC phone service? Nada. Pocket and CricKet were dead in the water as well.

Oddly enough, the Fredericksburg Standard Radio-Post cites Five Star Wireless as not being down during the incident, for local or long distance service. This on one hand makes sense, as Five Star doesn’t necessarily use phone-company T1 lines to power their cell sites like other providers do. However I tried calling a Five Star connected friend yesterday during the outage and couldn’t get through. I’m going to keep poking around for more info on what happened with that provider.

On the landline front, you could call to your own exchange but no further. This was the case for both Verizon and Windstream…and, as remarked before, HCTC. Time Warner Cable digital voice users apparently had the reverse issues; able to call out but unable to call folks in town…just like everyone else in the world. Sprint’s IP-based long distance also stayed online…it’s what powers Time Warner Cable’s digital voice service (at least for the moment…see this article) and I suppose runs on TWC’s network in the Hill Country.

Internet was more of a mixed bag. Time Warner Cable’s fiber route is different from that of the AT&T cable that was cut (near Waring), so their service, including service for the City of Fredericksburg, continued unabated.

Windstream DSL also continued without issue in Fredericksburg; my mom’s workplace uses them and was able to send an e-mail out about the incident. The phone lines on which the DSL was served couldn’t call beyond Fredericksburg, but internet connectivity did just fine. Oddly enough, Kerrville DSL didn’t weather the outage with as much gusto, according to reports by the Kerrville Daily Times.

Verizon DSL was flat-out down. Bee Creek Communications, riding on that AT&T fiber for their backbone, was out as well. On the flip side, Hill Country Telephone Cooperative’s dual internet links (one to the north, one to the south) didn’t miss a beat. As such, Kerrville’s school system was also unaffected; they use HCTC.

The impact of the outage went a bit beyond calling from point A to point B. Gas stations and Fredericksburg’s Wal-Mart could not take credit card transactions. I’m guessing Wal-Mart has a T1 internet connection, and their T1 was down along with everyone else’s (Sprint and AT&T’s cellular backhauls, for example). The gas stations were probably using plain old telephone lines to process plastic. For some businesses, old-style credit card imprint machines were dragged out and used to process transactions, but for many businesses those machines just don’t exist, relegating purchases to cash or check only.

Last but not least, some banks stopped early for the day (the outage started around 3 p.m. Central Time), as they also tend to use T1 connections to get online. Other banks just processed transactions manually, following procedures similar to blackout protocol.

This just goes to show how much the Hill Country’s communications infrastructure relies on a single provider. Pretty scary if you ask me. It also shows that voice over IP, combined with a real redundant network, is just a good idea; with VoIP it doesn’t matter if you’re running on cable, DSL or mobile broadband, as long as you’ve got access to the Internet. It’s fortunate that the service came up when it did, that TWC, HCTC internet and Windstream DSL customers were apparently not affected, and that emergency services were quickly restored in the area. However running all communications through a single AT&T fiber link (as opposed to a ring, or at least some sort of redundant route!) is just plain bad policy.

Comments/thoughts/additions? Post ‘em!

Tweets compiled about the event (more coming):

Windstream – @bmodisett @kerrvillegossip yes there is an outage; currently no ETA ; fiber cut by another carrier^MC
HCTC – Another good reason to be on Hill Country’s Ethernet backhaul service! Listen up wireless carriers! Can you spell “Diverse Routes”. We can! Call me!!! http://bit.ly/bpPHZU
KerrvilleGossip - Phone service down,, Windstream to blame…. working on the story, looks like HCTC is about to be the BIG DOG! #kerrville (retweeted by bmodisett)\
KerrvilleGossip - Cell service is starting to come back! GO KERRVILLE!

bmodisett - what is up with #Kerrville and the cell phones, it was like they pulled the plug for everyone here.
MrAdkins - Any 3G outages going on out there in twitterland?
tsampson - Anyone in Kerrville experiencing cell phone/land phone outages? Please at reply for potential story on this crazyness.
bmodisett - @tsampson i guess windstream finally had it and are leaving town.
tsampson- @shueytexas Working theory is that Aliens have landed, sending out a electromagnetic pulse… or something to that effect.
tsampson - @mradkins Also, have you heard anything from any carriers? I went over to Windstream and Verizon. No cause yet.
mradkins - @tsampson Just in from Sprint: there is an AT&T Outage affecting both cell and landlines west of Boerne on I-10 and north to Dallas.
mradkins - @tsampson But not sure if this is true. One carrier will blame another. WS won’t say they have a problem even if you drop by.
tsampson - So it’s Apocalypse Cell in the Hill Country today. Read all about it in the Daily Times’ team coverage tomorrow.

KENS5 - Communications cut in Texas hill country, services impaired http://bit.ly/b0QhIz
KVUE - Communications cut in Texas Hill Country, services impaired: It’s a terrible Tuesday for some people living in the… http://bit.ly/bDnGdc
News4WOAI - Major communication outage in Hill Country area: News 4 WOAI has confirmed that a fiber line was cut somewhere bet… http://bit.ly/awvd1X (retweeted by barbaramagana)
KABBFOX29 - Cut Fiber Line Causes Large Outage in Hill Country: A contractor working on roadside drainage causes a major…
http://tinyurl.com/yh3fku5

phlezk - DUDE WTF? phones (cell AND LANDLINE) in the Hill Country (my home!) are DOWN?! HOW THE–?!
phlezk - News 4 WOAI confirmed a Verizon fiber line was cut somewhere between Fredericksburg & Comfort, causing outages throughout the Hill Country.
that1collegeguy - @phlezk Dude, the entire cell network is like…non-existent at the moment!
that1collegeguy - NEWS FLASH: Texas Hill Country economy collapsing rapidly due to cell & landline networks crashing. LifeAlert reports record “pushes”!
BobGambert - Anybody up in the Hill Country? Talk to me about the major comm outage.
bbridges620 - AT&T says it is Verizon’s fault phone are down in the Hill Country, Verizon says it’s AT&T. What is up with Red Vs Blue everywhere? (retweeted by Brandy_Williams)
jfstx - all verizon communications down in Hill Country, no cell, home service , sprint is limping along on 1xrtt
priskillya - the entire hill country was cell phone and land lineless today. i also had no internet. that’s ok cause i had to move. it was just a pain
katiewright - cell service is finally FINALLY back up after about seven hours of being down in the entire Hill Country area.
2dogexpedition - phones in hill country down someone cut a fiber optic
metroane - this fn blows…….major phone, cell, internet services out since 1pm in hill country, can’t upload mag to press……

TaxesNews4u – San Antonio News: Widespread phone outage in Hill Country: Kerr, Gillespie counties are among the areas affected http://bit.ly/aW1t1y
SanAntonioCP - San Antonio: Major communication outage in Hill Country area http://bit.ly/dw8pUi
tweetsa - Widespread phone outage in Hill Country http://bit.ly/ayMHTA
salocalnews - Major communication outage in Hill Country area – http://fx.ly/a/16_78e8750bdc
salocalnews - Widespread phone outage in Hill Country – http://fx.ly/a/16_ad169f14d2
MsTechAngel - Widespread phone outage in Hill Country: By Zeke MacCormack – Express-News Damage to an AT&T fiber-optic line caus… http://bit.ly/cwMjyH
JeffVaughn - Breaking: Parts of Hill Country experiencing phone outage, including #Kerville: http://tinyurl.com/ydvwm2m
wrstech - Major communication outage in Hill Country area http://shar.es/mgfnZ
wxProBrian - Red vs Blue! RT @ksatnews: Now, Verizon tells us it’s an AT&T Fiber line that was cut – causing phone/cell/net outages in Hill Country.
GSMTalks - GSMtalks :: Service returns after widespread phone outage in Hill Country http://bit.ly/bml2pm
SanAntonioLocks - Phone Lines Out In Hill Country: A severed AT&T fiber line has cut cell, land and Internet connections in the Hill… http://bit.ly/b8Q9cj

1. SIP VoIP, over 3G and WiFi, with Bluetooth and in the background
I’ve successfully run phone calls over AT&T’s HSPA network, so assuming the processor on the iPad is “all that” this shouldn’t be difficult. If I can get a SIP client that connects to a Bluetooth headset and allows other applications “up front” without killing voice quality then I’ve got a very versatile system for nformation gathering while “on the phone.” This is particularly useful when you see my other requests below:

2. Full-fledged RDP, SSH and VNC clients…and either GoToMyPC or LogMeIn if at all possible
The iPad can be an excellent remote access tool; its screen has enough resolution to run any modern OS without having to scroll around. Add in the keyboard accessory and you’ve got a lovely little remote terminal that works anywhere AT&T has service (I’m talking about the 3G version here). Or any other non-AWS GSM or 3G carrier for that matter; the device is unlocked. SSH would be a godsend for doing command line heavy lifting, though you’d need a host to SSH to in this case (not a big deal for me). Even nicer, all of the above will likely be available for the iPad as App Store apps; no jailbreaking required.

3. A high-quality PDF viewer
I want a PDF viewer that can search and copy text, preferably at least as good as Preview. Yes, Preview can be outdone with PDFs…it’s not the most compatible piece of PDF reader in the world unfortunately.

4. File download capability, preferably with an FTP/SFTP client as well as your standard HTTP access
The iPad has an accessory port, and that accessory port already allows for an SD card reader and a USB port. If you could mount that USB port or SD reader as a drive letter (possible with the OS, but probably will need to be hacked) then there’s a TON of stuff you could do with the iPad, though things start to get kludgy once you realize that you can’t hook up a keyboard and a hard disk/SD card simultaneously. Hmm, maybe the keyboard dock could have a pass-through dock connector…or have an application that would use internal memory as a cache for editing files so you could unplug the drive, edit the files with the keyboard attached, then plug back in and write the files. Heck, wile I’m at it, why not let shares on AFP and SMB be mountable to the device?

5. WiFi printing
This will likely be implemented in short order as an application, though hopefully it will be available throughout the operating system. THat way I can go to a website, grab a PDF of some sort (documentation, textbook, whatever) and send it to a networked printer. No muss, no fuss.

6. Network diagnostics, both standard and wireless (including 3G)
I’m not just talking ping/traceroute/speedtest, though those would be nice. I’m talking about field-test-level stats on wireless and 3G. Why 3G? Well, for one AT&T isn’t the only 3G carrier out there, and if network engineers on smaller carriers can troubleshoot their cellular deployments on a cheap-ish, small device then maybe there will be more field testing going on. On the WiFi side, an iStumbler-esque application would be lovely, complete with a signal strength history graph, with geotagging of signals if at all possible (though this would take a toll on battery life). Heck, add the geolocation to the 3G field test if at all possible and just for kicks port WireShark to the iPad and you’ve got one powerful network analysis device in a pound-and-a-half package.

7. Voice recorder, preferably with backgrounding
I’m guessing this will be included when the iPad ships, but it would be lovely to be able to take the iPad to a lecture or an interview and use it to record the conversation. Even cooler would be an application that would combine a text editor with the voice recorder; in “review” mode you could seek through a recording by going to a place in the text notes, and the other way around.

8. OneNote-like application
Capacitative styli are available. Make an application that allows for text notes, handwriteen notes/diagrams and even voice notes all together. Great for school and for college newspaper writers like myself

9. Access to a bash shell on-device
I’m sure this will only be available when the device is jailbroken, but it would definitely be useful. A full shell, complete with nano for text editing etc., would make the iPad a lightweight BSD workstation, and who doesn’t like lightweight *nix workstations?

Sure, I’m reaching a bit for the above things, but the hardware is all there; you just have to either open up the software or make an application. There are a few subsystems that would be nice to have,a dn they might even be available inthe upcoming iPhone OS 3.2. In all honesty, I think Apple is going to have to make this device rather hackable for enough people to buy it; they’re blazing the trail for a new category of device and if they shun the “Woz box” mentality they’ll be selling very few of these things once the fanboys and their contituents finish buying (which should take about 45 days).

Traceroute to Google:

traceroute to google.com (74.125.159.104), 64 hops max, 52 byte packets
1  64.sub-66-174-185.myvzw.com (66.174.185.64)  82.040 ms  71.681 ms  140.969 ms
2  127.sub-66-174-184.myvzw.com (66.174.184.127)  85.990 ms  88.710 ms  93.889 ms
3  146.sub-66-174-44.myvzw.com (66.174.44.146)  94.923 ms  99.735 ms  92.011 ms
4  235.sub-66-174-45.myvzw.com (66.174.45.235)  120.975 ms  88.777 ms  86.127 ms
5  194.sub-66-174-45.myvzw.com (66.174.45.194)  86.937 ms  99.731 ms  92.958 ms
6  98.sub-66-174-44.myvzw.com (66.174.44.98)  89.816 ms  92.702 ms  91.968 ms
7  6.sub-69-83-96.myvzw.com (69.83.96.6)  88.964 ms  90.749 ms  101.982 ms
8  3.sub-69-83-96.myvzw.com (69.83.96.3)  91.973 ms  89.745 ms  92.877 ms
9  253.sub-69-83-97.myvzw.com (69.83.97.253)  86.946 ms  87.701 ms  86.078 ms
10  gigabitethernet3-0-0.gw1.dfw7.alter.net (65.208.8.17)  93.040 ms  86.646 ms  86.916 ms
11  0.ge-2-1-0.xl3.dfw7.alter.net (152.63.103.6)  99.833 ms  102.809 ms  89.928 ms
12  0.ge-6-0-0.br2.dfw13.alter.net (152.63.96.181)  89.939 ms
0.ge-4-0-0.br2.dfw13.alter.net (152.63.96.45)  91.661 ms
0.ge-6-0-0.br2.dfw13.alter.net (152.63.96.181)  89.622 ms
13  te-5-1-0.edge2.dallas3.level3.net (4.68.111.121)  99.987 ms  116.649 ms
te-5-2-0.edge2.dallas3.level3.net (4.68.111.173)  90.885 ms
14  ae-83-80.ebr3.dallas1.level3.net (4.69.145.180)  92.662 ms
ae-73-70.ebr3.dallas1.level3.net (4.69.145.116)  92.622 ms
ae-63-60.ebr3.dallas1.level3.net (4.69.145.52)  97.740 ms
15  ae-7.ebr3.atlanta2.level3.net (4.69.134.22)  115.793 ms  137.608 ms  130.824 ms
16  ae-11-51.car1.atlanta1.level3.net (4.68.103.2)  120.070 ms  117.856 ms  116.900 ms
17  google-inc.car1.atlanta1.level3.net (4.78.209.194)  133.861 ms  118.900 ms  117.002 ms
18  209.85.254.241 (209.85.254.241)  122.982 ms  116.700 ms  119.880 ms
19  209.85.254.2 (209.85.254.2)  113.964 ms  120.700 ms
209.85.254.6 (209.85.254.6)  137.883 ms
20  yi-in-f104.1e100.net (74.125.159.104)  109.804 ms  129.913 ms  131.328 ms

Traceroute to SoftLayer (specifically, the server that hosts this website):

traceroute to echo.mddhosting.com (74.86.48.34), 64 hops max, 52 byte packets
1  64.sub-66-174-185.myvzw.com (66.174.185.64)  73.485 ms  72.917 ms  73.973 ms
2  127.sub-66-174-184.myvzw.com (66.174.184.127)  91.918 ms  94.745 ms  184.212 ms
3  146.sub-66-174-44.myvzw.com (66.174.44.146)  100.537 ms  99.876 ms  95.165 ms
4  235.sub-66-174-45.myvzw.com (66.174.45.235)  94.993 ms  100.098 ms  98.174 ms
5  194.sub-66-174-45.myvzw.com (66.174.45.194)  98.010 ms  99.438 ms  93.713 ms
6  98.sub-66-174-44.myvzw.com (66.174.44.98)  94.994 ms  89.077 ms  85.875 ms
7  6.sub-69-83-96.myvzw.com (69.83.96.6)  100.961 ms  105.189 ms  88.124 ms
8  3.sub-69-83-96.myvzw.com (69.83.96.3)  87.943 ms  96.379 ms  97.863 ms
9  253.sub-69-83-97.myvzw.com (69.83.97.253)  98.888 ms  68.155 ms  87.944 ms
10  gigabitethernet3-0-0.gw1.dfw7.alter.net (65.208.8.17)  86.948 ms  69.442 ms  96.147 ms
11  0.ge-2-1-0.xl3.dfw7.alter.net (152.63.103.6)  112.539 ms  89.027 ms  85.911 ms
12  0.so-6-0-0.gw3.dfw13.alter.net (152.63.100.73)  85.015 ms  91.726 ms  86.006 ms
13  internapgige1-gw.customer.alter.net (65.208.15.230)  113.416 ms  96.205 ms  98.848 ms
14  border3.tge3-1-bbnet1.ext1.dal.pnap.net (216.52.191.22)  125.890 ms  95.198 ms  74.901 ms
15  te2-1.cer03.dal01.dallas-datacenter.com (216.52.189.30)  93.058 ms  94.118 ms  95.968 ms
16  po3.dar02.dal01.dallas-datacenter.com (66.228.118.211)  99.913 ms  100.785 ms  107.881 ms
17  po2.fcr02.dal01.dallas-datacenter.com (66.228.118.182)  111.010 ms  92.603 ms  121.966 ms
18  echo.supportedns.com (74.86.48.34)  97.877 ms  96.688 ms  153.805 ms

So pings are a bit lower, on average, than either AT&T or Sprint. They’re also rather predictable, unlike those of CricKet. Then again, I have full EvDO signal sitting at the dining room table with the USB727′s antenna extended. For what it’s worth, this traceroute is pretty close, after the MyVZW hops, to what I see on the family’s brand-spanking-new DSL connection, though that can be expected since both services are from the same company.

Here’s a speedtest…

Well gosh, at least at my location Verizon is nearly as fast as AT&T on the downstream, respectable on the upstream and a good bit more enjoyable to use than AT&T. Heck, the EvDO connection soundly beats the new Verizon DSL, though the DSL connection doesn’t cost $60 per month, nor does it have a ridiculously low data transfer cap.

That said, I’m not the only one on the tower…

Still respectable, just not quite as nice.

Now for a ping test…

Right in line with what I’ve seen on Sprint EvDO, except with a little less latency.

The bottom line: Verizon knows how to run a solid mobile broadband network. So does Sprint. Whether one serves you better than the other depends on who has the big pipes to the towers in your area, and who has the towers in your area period.

Well, and there’s the question of whether you want a 10GB cap for $60 or unlimited service for $70, both via Millenicom of course…unless you really, really want to sign a contract in exchange for a free EvDO modem.

Which provider would I choose at this point, if I had to pick up mobile broadband again? Tough decision; I own an AT&T aircard and don’t want to sign a contract, so the other options have an added expensve for me. With that in mind, I might end up going with AT&T for a month or two, but Sprint if I needed a home replacement connection. If I didn’t need to use the connection as my only internet and didn’t have an AT&T card already, I’d likely choose Verizon, with Sprint as a very close second.

At this point though, for someone new to mobile broadband, there’s one company I wouldn’t recommend anyone to go with, at least for my area: AT&T.

So naturally I pulled out my HTC Touch Pro, which has WiFi, to see what network they were running on. After registering for an account on the hotspot provider (Less Networks, never heard of them) I checked whatismyip.com to see who I was on. Sure enough, the connection there was GVTC, registered as “Guadalupe Valley Telephone”. Right before we pulled out of the place I opened up my laptop, logged into the WiFi network that way and did some tests. Here’s what I got for a traceroute:

traceroute to softlayer.com (66.228.118.53), 64 hops max, 52 byte packets
1  192.168.169.1 (192.168.169.1)  194.115 ms  181.762 ms  77.970 ms
2  96-8-192-3.block0.gvtc.com (96.8.192.3)  94.851 ms  107.971 ms  146.974 ms
3  12.88.219.169 (12.88.219.169)  47.516 ms  218.308 ms  149.762 ms
4  cr1.santx.ip.att.net (12.123.154.10)  81.998 ms  58.617 ms  64.528 ms
5  cr2.dlstx.ip.att.net (12.122.30.130)  71.363 ms  55.947 ms  70.918 ms
6  gar6.dlstx.ip.att.net (12.122.100.65)  52.436 ms  58.138 ms  56.583 ms
7  12.90.228.14 (12.90.228.14)  52.971 ms  48.888 ms  80.834 ms
8  border3.tge4-1-bbnet2.ext1.dal.pnap.net (216.52.191.83)  52.188 ms  51.494 ms  52.917 ms
9  te2-1.cer03.dal01.dallas-datacenter.com (216.52.189.30)  55.883 ms  54.237 ms  55.133 ms
10  po3.dar01.dal01.dallas-datacenter.com (66.228.118.209)  53.084 ms  84.979 ms  47.910 ms
11  po1.slr01.dal01.dallas-datacenter.com (66.228.118.138)  66.371 ms  54.266 ms  57.160 ms

I have a feeling that the WiFi conection was flakey, otherwise I would have gotten a better traceroute.

Speed-test-wise I got the following:

I’m guessing the line was provisioned at 8 Mbps down and 1 Mbps up, but I can’t be sure.

The IP address, for what it’s worth, was 96.8.236.103.

All in all, not a bad hotspot, and the traceroute is nice and short between where I was and AT&T’s network in San Antonio, which GVTC appears to use exclusively to connect to the internet nowadays. So someone using an AT&T aircard and GVTC FTTH will find themselves with similar traceroutes within just a few hops, at least around here.

I guess I’ll leave it at that…if I talk any more about GVTC I’ll start getting envious of their 20/3 and 40/10 FTTH service, which I have no chance of getting at my place. Can’t even get DSL there :/

One problem arose when I got home: my aircard kept switching to EDGE during the middle of a data session. Not a good thing for a $60 + taxes and fees data plan. So I needed to force 3G on the card. The problem there was that only the Windows Sierra and AT&T clients allow forced 3G. So (not knowing that settings are stored on the modem and thus once I set 3G-only on my Windows computer it would be 3G-only on my Mac) I scrounged around and found this page. I followed those instructions and was able to make sure my aircard was 3G-only.

If you don’t want to read through a (relatively short) page on how to kick your aircard into high gear, you’re in luck. Here’s what you’d do on a Mac to force-3G a Sierra aircard. Note that this procedure was performed on OS X 10.6.1 and with the Apple “native” driver installed (a tiny download from here). This is also with the latest modem firmware, which I had to use the Sierra driver (vs. the native one) to upgrade to.

1. Go to Terminal (Applications > Utilities)
2. Type in screen /dev/tty.wwanCMD 115200 and hit enter.
3. Type in AT!band? and hit enter.
4. If you see the number 02, you’re already locked into 3G-only mode. Otherwise, type in AT!band=02 and hit enter to set the mode. If you wanted to go 2G-only, the code would be 04 instead of 02. AUtoselect is 00.

Hope this helps! Now I’m cooking along with 2-3 bars of signal (-87 dBm right now) and pulling down these speeds:

Though connection quality is quite a bit more variable than Sprint’s (slower, particularly on uploads) connection at my place…

In comparison, SPrint usually pulls down 1.3 Mbps down, 300-500 kbps up and gets a B on connection quality.