American expat in Switzerland here. Using Speedtest.net I get 246.08/15.21 Mbps. I pay the cable company the equivalent of $98 USD/month for 250/15 internet service (no data caps) and cable TV (my wife likes watching US sports, so we have the "all-inclusive" TV package that includes some US sports channels). I originally had the 35/5 plan, but upgraded to the 150/10. They discontinued that plan and switched me to the 250/15 plan, which was only $5/month more.

If I wasn't satisfied with them, Swisscom (major telco) and the electric company each offer fiber-to-the-home, with up to 1000/100 speeds and no caps. There's other options for DSL too, but not nearly as fast.

Comcast, a major US ISP, has a comparably-priced plan that goes from $89/month for the first year to $119/month for the second year and then up to $148/month thereafter. They offer a bunch of TV channels and 25 Mbps internet, plus data caps. That's absurdly awful.

As an American, I find it ridiculous that wholesale bandwidth in the US (e.g. connectivity in a datacenter) is dirt cheap and fast (as an example, Hurricane Electric offers 10GigE transit for $0.45/Mbps) but that retail bandwidth available to end-users is so expensive, slow, and limited by data caps and the like. Things really need to change.

Symmetric and asymmetric keys are different things and have different key lengths. One cannot directly compare key sizes between two wholly different classes of ciphers. There are numerous reasons, mostly involving arcane mathematics, why asymmetric ciphers require longer key lengths than symmetric ciphers to offer similar levels of protection.

For example, a 1024-bit RSA key (RSA is an asymmetric cipher) is essentially equivalent to an 80-bit symmetric key (AES, 3DES, etc. are symmetric ciphers). SHA1, a hashing algorithm, provides less than 80 bits of security; those wishing stronger signatures are switching to SHA-256 (which offers 128 bits of security) and SHA-512 (which offers 256 bits).

A 2048-bit RSA key, such as those used by most CAs and web servers these days, has the same strength as a 112-bit symmetric key. NIST says they should be good enough until around 2030.

3072-bit RSA keys offer the same strength as a 128-bit symmetric key. A whopping 15,360-bit RSA key would be needed for 256-bit security; the same level of security could be achieved with a 512-bit elliptic curve key, which would be much, much faster than such a large RSA key.

Use a stateful firewall? NAT is not a firewall.

Just because something has a globally unique IP address doesn't mean that it's globally reachable.

Who says you need to tell the truth on those questions?

Q: "What is your mother's maiden name?"
A: "Purple monkey dishwasher."

Of course, you should keep a record of those questions and answers so you can correctly answer them if the need arises.

(Emphasis mine.)

Why not? That's basically what Jon Postel did: he basically singlehandedly administered the DNS root and was IANA.

Sure, things are different now, but we certainly have had random people running the internet. It worked then, why not now?

Companies like what? The ones making network-tapping hardware and whatnot cater toward a limited market, not the general public. Certificate authorities directly transact with server administrators, but their primary audience are end-users and they have wide public exposure. If a CA was found to be doing shady things, browsers would remove their roots. That'd basically kill off the offending CA.

Why not? There have been roots that have been revoked due to being compromised and which have issued bogus certs (e.g. DigiNotar). That's caused some chaos, but people adapted.

Sure, VeriSign is large and commands (either directly or through its subsidiaries) a substantial fraction of the CA market. Nuking it would be a Very Big Deal that browsers wouldn't take lightly, but I have no doubt that if it were shown that VeriSign (or Comodo, or other CAs) were found to be issuing bogus certs for the government to compromise people, they'd get their roots pulled by browsers. That's a death sentence for a CA, hence my skepticism in response to the proposal that they're actively assisting governments.

Agreed. That would be nice.

Sure, they could, but I doubt they are.

If VeriSign gets caught issuing bogus certs for the government, browser vendors will revoke their roots. That's basically a death sentence to companies like VeriSign (rather, their cert-issuing division).

While typical users won't notice, there's still plenty of risk to getting caught, particularly when targeting anyone using major web properties: Chrome, for example, has a bunch of high-profile sites "pinned" and will report back to Google if bogus certs are being used (they identified a bunch of MITMing with compromised certs in Iran in this way). Other add-ons like Perspectives make it easier to detect if unexpected certs are showing up.

Could they get away with issuing infrequently-used certs for highly-targeted, one-off uses? Possibly, but each time they do the risk to their entire business increases.

I suspect the government would much prefer to do things sneakily in the shadows, rather than involving major CAs in such a risky role.

Perhaps. Depends on the location and provider. Here in Bern, Switzerland, the cable company offers 250/15 internet for CHF 89/month ($98 USD). That's only $10 more than the 200/200 for $89.95 offering. Not unreasonable. For CHF 105/month they package a bunch of cable TV channels (including European and American sports) and 250/15 internet.

Swisscom, the incumbent phone company, has fiber-to-the-home. 300/60 internet with even more TV channels costs CHF 154/month. They offer up to 1000/100 connection if you're willing to pay and extra CHF 80/month above the CHF 154 rate. That's USD $258, only $8 more than the 1000/1000 plan offered in Clarksville (though the Swisscom offering does include TV. Phone is an extra CHF 15).

Then again, Switzerland does tend to be expensive. You may get cheaper service in other countries, but it is quite comparable in terms of cost here.

True, but how often does one need to revoke a certificate? Other than Heartbleed, I think I've only revoked one certificate in the last 10 years or so. Amortized over that timeframe, the costs are negligible. That said, I would like it if StartSSL would offer free revocations in the case of something like Heartbleed, where certs are compromised through no fault of the customer, but I understand the business reasons for not doing so (CRL/OCSP isn't free).

Of course, I've abandoned several certs where I deleted a VM hosting a site I no longer needed, but since the cert was not compromised and the private key was deleted, I just let the expiration timer run out. No big deal.

Why not just use SNI? I have multiple SSL-enabled virtualhosts running on a single server, and other than Internet Explorer on Windows XP and Android 2.x (neither of which I care about, as the former is EOL while the latter is effectively EOL) every browser on desktop and mobile devices works properly. I spend more money every two weeks on caffeinated beverages than my entire annual budget for SSL certs, and I have more than most. My cert budget is dwarfed by hosting costs (which I pay regardless of SSL support or not).

If you don't care about those systems (and I don't), SNI is perfectly satisfactory. If you need to support those old systems for some reason, you're probably a commercial enterprise who can afford IP-based SSL or wildcard certs. For typical individuals or small/medium-sized organizations using SNI, adding SSL support to your sites will essentially be a non-issue in terms of cost.

If anything, Google adding a small boost to SSL-enabled sites should encourage and improve support for SNI and hopefully sweep away the few older browsers that don't support it. I'm all for it.

Seconded in regards to the N66U. It's a fantastic router. I've been running Tomato Shibby for years (most recently v121) and it's been rock-solid, reliable, and stable.

There's only one downside: Tomato doesn't include the necessary kernel module for hardware accelerated WAN-to-LAN NAT/routing. This only matters if your downstream WAN bandwidth is greater than ~120Mbps. If your downstream bandwidth is less, the software routing can keep up and you'll run at full speed. If your downstream bandwidth is greater, you will be limited to ~120-130Mbps, as that's as fast as the N66U can route in software. LAN-to-LAN bandwidth will run entirely in hardware regardless of what firmware you have.

My ISP just upgraded me to a 250Mbps downstream link, so I reluctantly went back to the factory firmware to take advantage of the hardware acceleration. It's clunky and annoying compared to the elegance of the Tomato web interface, but it works. The Merlin firmware maintains the look-and-feel of the factory firmware, includes support for hardware acceleration, fixes a few bug and adds a few features (but not as many as Tomato) that makes it suck less.

I highly recommend the N66U.

A validity time of one year is pretty standard for SSL certs (paid certs often charge per year). Could they issue them for 20 years? Sure, but a one year validity is not unusual. Class 2 certs are good for two years.

I get renewal notices two weeks prior to expiration. That's pretty reasonable. If I recall correctly, I can generate a new cert for my site any time in that two-week period, so I don't need to wait for the cert to expire before replacing it.

While I wish they allowed free reissuance of certs at any time, I don't really see why requiring revocation is a showstopper.

That's also the case for pretty much any of the inexpensive paid certs too. You can always get a wildcard cert but most CAs charge at least $100/year for a single wildcard cert. StartSSL charges $60 for Class 2 validation, and you can issue unlimited certs (wildcard or not). Organizations can get Class 2 certified for $120 ($60 for identity verification, $60 for organization verification) and can issue unlimited certs. For a company needing more than one cert, StartSSL is still cheaper.

Considering their paid certs are often cheaper than comparable offerings from other CAs, it doesn't really seem unreasonable to me. Doubly so because they're run by competent people who respond promptly to inquiries, even from free users. I've been a StartSSL customer for years (and also used other CAs like GoDaddy, Comodo, Thawte, etc.) and the customer service from StartSSL has always been excellent.

If you don't want to get a StartSSL cert or they don't meet your needs, that's fine. NameCheap and others sell single-domain Comodo certs for $9/year. RapidSSL certs are a buck or two more per year. That costs less than a single beer at the local bar. Hardly a massive expense.

Quite the contrary: StartSSL is accepted by every major browser and SSL/TLS library, and has been for years.

Well-known sites, like EFF.org, LibreOffice, and others use StartSSL-issued certs and don't have any issues. Sure, they're not Google-sized sites, but they're fairly major.

That seems to match with what I'm saying: when the phone is constantly searching for signals it has the receiver enabled all the time and the gain turned up to maximum, using more power. When it is in an area of low-but-there signal, the receiver isn't powered up as often, but the gain is still high, so it uses a medium amount of power. When you're out in the open and there's lots of signal, the receiver isn't powered up as often and the gain is low, so it uses the least amount of power.

I apologize if I wasn't clear before.