Verizon Slowing Linux Access to Website?
Once you sign in to your account, the Verizon Wireless website is slow to the point of being unusable on Linux. It’s also slow to the point of being unusable on Windows XP, at least in a virtual machine. And I don’t mean I’m being impatient. I mean that I have to wait multiple minutes to load a single new view (i.e. part of a page), and that I sometimes get browser timeouts.
I tried multiple times over multiple weeks. I tried in both Firefox and Internet Explorer (the latter only from my Windows VM). I Googled around and saw some others (but curiously, not very many) complaining of similar issues.
So I gave up on trying to use the site. Which is unfortunate, since it has a lot of useful features. As a (now former) US Cellular customer with a dumbphone, I had no desire to use a website for anything (if they even have one). As a Verizon customer with a smartphone, it would be really nice.
I did try to use the site to contact customer service, to report the problem, but I couldn’t load the page to do so.
Googling about the problem again, I came upon this forum post (emphasis mine):
I thought it was because I use AT&T for internet. lol
Same issue here though, slow as crap on FF and Linux and even in Vbox running XP on Linux.
I can fire up my laptop running XP and get right into the Verizon site.
I don’t understand how it’s even possible that a site could load slow unless you’re running Windows?
It’s true. The Verizon Wireless website is balls slow when accessed from Linux, even using Internet Explorer in a Windows XP virtual machine. It is lightning fast when accessed from a computer running Windows XP natively.
I told my officemate this bizarre story. He says that there actually are hacks, such as looking at MAC address ranges, that can give websites useable clues about whether or not requests are coming from a virtual machine.
Do we have a conspiracy on our hands? Maybe, maybe not. I recently saw complaints about Verizon doing an on-air update for many of their customers’ phones, making Bing the default browser. Seems possible that Microsoft and Verizon are “scratching each other’s backs.” What Verizon is getting in return is up for speculation. OTOH, I’m not saying I really think there really is a conspiracy—it could very well be a bizarre technical issue. I think the issue deserves further investigation.
For the most part, I’ve been very happy with Verizon. I’ve only had a few issues.
- The topic discussed here
- The topic discussed in my previous post
- Visual Voicemail “ineligibility” (I have a pending customer service inquiry about this)
All things considered, Verizon has done a lot of good: they sold me a great smartphone and have excellent coverage. AFAIK only US Cellular can compete with them in the latter category in my area, and US Cellular lost my business after many years for not satisfying in the former area.
Well, they aren’t telling if you’re in a VM by your MAC address, since that’s only locally significant. In fact, if you are using NAT for your VM network access, no one outside of your computer could tell (at least by MAC address).
Now, they could be doing OS fingerprinting at their firewall to tell what OS you’re running. If you’re running your VM in NAT mode, then it’s still using the linux kernel IP stack, so it would still be fingerprinted as linux. Try putting your VM into bridged mode, then access it from the VM.
If the above works, I’d imagine it’s a strange firewall issue, and has nothing to do with the web server or coding. Would still be rather strange, to do OS fingerprinting on a site as large as Verizon’s would take significant processing power.
How does this fingerprinting work, in a nutshell?
Care to speculate on whether this is intentional or not?
Either way, having a website that *does not* work with Linux is, IMHO, pretty unacceptable — they should be testing their site with Linux. Of course, it *could* be something even more specific to my machine than the OS. I would like to see some other Linux people try this.
Oh yes, and a friend of mine has the Nexus One….it’s awesome. If/when they make a landscape hardware keyboard variant for CDMA, I’ll be all over that!
They already have one, it’s called Motorola Droid.
Perhaps, but it doesn’t have the 1GHz Qualcomm Snapdragon processor – that thing flies. It’s also ugly as sin.
What is ugly as sin? The Snapdragon, the Nexus One, or the Droid?
Anyway, I don’t think the Droid is ugly in the least… in fact, I think it looks better than most phones out there. I don’t like things to be overly stylized. I’m not trying to impress a 9 year old. I would make the exact same comments about the aesthetics of my ThinkPad.
Also, I can’t think of a good reason to need a 1GHz CPU on a smartphone, unless you want to waste power and run Windows. (Today. Yes, I realize this remark is in the same category as “nobody could ever want more than 24k” and the like.)
That was a little ambiguous – the Droid is ugly as sin. The snapdragon is a processor, I’m not really concerned how it looks, it’s fast as shit. The Nexus One looks OK, but there are a few things about it I dislike.
My biggest gripes with the Droid is the keyboard layout and that stupid d-pad beside it. I played with one in a Verizon store not too long ago, and I didn’t really like the look or function. Function would certainly improve with practice, though. And yeah, I don’t want an apple-looking phone, but I like my phone to be visually pleasing.
I just don’t agree that the Droid is ugly.
At first I thought I would agree with you about the d-pad. Honestly, it is useful when you’re trying to click on links on websites that are smushed together, which is pretty common. You really need that to select which link you want (as opposed to selecting one you don’t want).
Also, for me, the keyboard is quite comfortable for sending text messages, which is all I’d want to use it for. My fingers are probably average size though, while yours are pretty big. So, mobile keyboards are not one size fits all. Anyway, it definitely beats all the phones without hardware keyboards, and probably all phones with hardware keyboards except the phone you have.
BTW, I edited my previous comment a bit, so you may not have seen the whole thing.
A few additional comments to your previous points:
I generally love Thinkpad asthetics. I think the real difference, to me, is I like the Thinkpad lines, while I really don’t like the Droid lines. In an unrelated note, the VP of design for Lenovo has an excellent blog, http://lenovoblogs.com/designmatters/ . I need to add it to my feed reeder, but it’s generally very interesting. I started reading it for the articles about the X300 design. You might want to check it out in your free time (haha).
As for the 1GHz remark – well, Windows Mobile 6.5 is, for sure, very “bloated”. But, comparing Brett’s Nexus One to my Touch Pro 2 and a Droid, his phone is noticeably faster. The Droid isn’t too bad, and the TP2 can be horrible at times. I’ve learned to live with it, but the “snappyness” of the Nexus One was amazing to me.
As far as keyboards – I’m well aware I’m being spoiled by the excelent keyboard on the Touch Pro 2. I’m very worried that I won’t find one nearly as good in a few years when I’m looking for another phone. I would say the Droid keyboard is better than many others I’ve seen.
For links – I’ve gotten used to using my fingernail, or zooming the page in. I’ve also got arrow buttons, but to use those, I’d have to slide the keyboard out.
Response to OS fingerprinting on the way…
Thx for the blog recommendation. It’s been mentioned to me before (probably by you) but wasn’t in my feed reader for some reason.
Haha, don’t kill yourself on the OS fingerprinting thing.
Standard Verizon.com works fine on my machine – Ubuntu 8.04, using Firefox. Only issue was having noscript running (which you should consider using, if you don’t already), easily solved.
From what I understand, there are 2 forms of OS fingerprinting, Passive and Active. Passive, what would be used in this case, basically uses information it gleans from the IP and TCP headers in the packets it receives. Stuff like TCP sequence numbers and flags. There is also active OS fingerprinting, which involves sending strange packets to a host, to see what it returns. See my blog post, http://glimpseintoentropy.blogspot.com/2009/10/iptables-stealth-scan-detection.html – it briefly discusses scans that send TCP packets with malformed flags. The scans work off of what data is returned by the OS. My post discusses making the linux firewall, iptables, reply or not, based on what you want the scan to reveal.
Now, if you’re using a hardcore firewall, like OpenBSD’s PF, then it can take care of most of this for you. pfSense is a great firewall package, that has the power of pf and FreeBSD, with the simplicity and web-interface of a standard home router. This firewall can block all OS fingerprinting, both active and passive, as well as stop SYN flood DOS attacks in their tracks.
Intentional or not:
Well, this may be a moot point, since it works fine for me. I’d guess it’s unintentional – probably a network engineer thinking they’ve found a clever way to block some sort of attack. You may want to run wireshark on one of your sessions with Verizon.com and see what it reveals.
Try your VM in bridged mode and tell me what happens.
Thanks for the info on OS fingerprinting. Makes sense.
I think comparing verizon.com to the logged in user experience is apples to oranges. My problems only happen after I log in to “My Verizon”.