Live data from Hacker News

I fought my ISP's bad behavior and won

erichelgeson.github.io

71–80 of 115 posts

Re: I fought my ISP's bad behavior and won

#71

Earlier quoted context omitted.

Super shady stuff. I never rely on any ISP provided DNS servers Doesn't that mean that you don't benefit from nearby CDNs? Perhaps worth the tradeoff anyway.

8.8.8.8 and 8.8.4.4 are anycast.

I don't see how this necessarily helps.

Edit: I did a little research on CDN resolution back in undergrad. One of the things that was most difficult to measure, or even define, was DNS "performance."

Just because Google anycasts its DNS network does not guarantee the best performance. Running namehelp[1] on many machines I often found that the "fastest" DNS server was not in the obvious set of DNS servers that you use.

So I guess the answer to the parent's question is: yes, possibly, but it's hard to define what "fastest" even means here. I would say that jrockway's response is even incorrect in his (implied) definite assertion that it cannot degrade your experience. If you want to know, you should measure from your endpoint. There is no other way.

[1] http://aqualab.cs.northwestern.edu/projects/namehelp

Re: I fought my ISP's bad behavior and won

#72

Earlier quoted context omitted.

And GOOG-owned. Spam from a small-scale ISP, or tracking by a big ad firm? You lose either way.

https://developers.google.com/speed/public-dns/privacy "We built Google Public DNS to make the web faster and to retain as little information about usage as we could, while still being able to detect and fix problems. Google Public DNS does not permanently store personally identifiable information."

"does not permanently store personally identifiable information"

If your employer gave a shit, this would read: "does not store information". But it doesn't, and they don't, and you're a cheap shill.

Re: I fought my ISP's bad behavior and won

#73
post #61

Cox does something similar but bypasses the the DNS records and just slipstreams in a response. I noticed Cox would redirect javascript requests to their own HTTP server and put in their own snippets, effectively doing mass javascript injection. The snippet ended up being some sort of alert about upcoming maintenance, but using a malicious technique for a benign purpose is the path to the dark side. Use HTTPS! (I use…

Rogers (Canada's Comcast) does the same thing, but for warning you about your bandwidth usage.

Re: I fought my ISP's bad behavior and won

#75

Super shady stuff. I never rely on any ISP provided DNS servers. I'm glad you talked to the the etailers to let them know what was going on. These business practices do introduce latency, regardless of what he told you. Not to mention, they are highly unethical and dishonest.

Super shady stuff. I never rely on any ISP provided DNS servers Doesn't that mean that you don't benefit from nearby CDNs? Perhaps worth the tradeoff anyway.

I think you would be hard pressed to find a public well known service that is depended on ISP CDNs.

Google? No. Games? No. JS libraries? No.

Even if some might use CDNs, they are likely to also use geo-aware DNS servers. It cheap, don't cost anything, and can be combined with CDNs on ISP level.

Re: I fought my ISP's bad behavior and won

#77

Super shady stuff. I never rely on any ISP provided DNS servers. I'm glad you talked to the the etailers to let them know what was going on. These business practices do introduce latency, regardless of what he told you. Not to mention, they are highly unethical and dishonest.

Super shady stuff. I never rely on any ISP provided DNS servers Doesn't that mean that you don't benefit from nearby CDNs? Perhaps worth the tradeoff anyway.

No, good public DNS servers (like Google's) forward your subnet to the resolver to ensure that you get correct responses: https://en.wikipedia.org/wiki/Content_delivery_network#edns-...

Re: I fought my ISP's bad behavior and won

#78
post #68

Eric, I am very sorry to see this happen to you. Unfortunately more and more companies are using our data for marketing purposes. All is not lost though. There are several ways you can protect yourself from these practices. The first thing I would do is get a router capable of using dnscrypt-proxy ( http://www.opendns.com/technol... . Then you can be confident that your DNS traffic is not being modified by your ISP.…

> It does require that you have trust in a 3rd party DNS provider like OpenDNS ...

The same OpenDNS that hijacks NXDOMAIN responses?

Re: I fought my ISP's bad behavior and won

#79
Here it goes: Behind a ISP-wide cache. Any 'traceroute' passes by transtelco.net (ISP used to have their own infraestructure for voip services Megafon) now i have 5/6? DNS jumps! and all my traffic going to Transtelco.

  traceroute to news.ycombinator.com (198.41.191.47), 30 hops max, 60 byte packets
  1  customer-GDL-**-***.megared.net.mx                 
I don't know how normal or ethic is this type of cache. No download limits, I have the 10mb and get 20mb(2000-2300kbps) downloads, for uploads is limited to 1mb.

Re: I fought my ISP's bad behavior and won

#80

Earlier quoted context omitted.

8.8.8.8 and 8.8.4.4 are anycast.

I don't see how this necessarily helps. Edit: I did a little research on CDN resolution back in undergrad. One of the things that was most difficult to measure, or even define, was DNS "performance." Just because Google anycasts its DNS network does not guarantee the best performance. Running namehelp[1] on many machines I often found that the "fastest" DNS server was not in the obvious set of DNS servers that you us…

Interesting, just tried out namehelp.

It told me my the fastest option for me was Dyn's servers, but those have unacceptable to me anti-spyware "safe mode" blockouts. #2 on its list was Google DNS which I was already using.

The graphs for HTTP performance said namehelp's choices were degrading my performance on average by 11ms.

Good app to add to the toolchest though. It'll be interesting to try it out when I travel.

Post reply on HN