January/February 2018 issue of acmqueue


The January/February issue of acmqueue is out now


Networks

  Download PDF version of this article PDF

ITEM not available

acmqueue

Originally published in Queue vol. 7, no. 10
see this item in the ACM Digital Library


Tweet



Related:

Yonatan Sompolinsky, Aviv Zohar - Bitcoin's Underlying Incentives
The unseen economic forces that govern the Bitcoin protocol


Antony Alappatt - Network Applications Are Interactive
The network era requires new models, with interactions instead of algorithms.


Jacob Loveless - Cache Me If You Can
Building a decentralized web-delivery model


Theo Schlossnagle - Time, but Faster
A computing adventure about time through the looking glass



Comments

(newest first)

Displaying 10 most recent comments. Read the full list here

Albert Klausevits | Tue, 02 Aug 2016 13:53:01 UTC

I had a trouble with the body's weight, back pain and also motion security. After physical medication recovery in New York Dynamic Neuromuscular Rehab (NYDNR) https://nydnrehab.com/ I feel wonderfull!!! Throughout DNS treatment, I treated for imbalances, disorders, problems with pose and control disorders with an approach that takes the specific back to placements of early growth as well as uses therapy to proceed function as it is endured. The training is conducted in the most all-natural (excellent) body placements. When learnt in this manner, the main movement mechanisms become automated offering basis for healthy.


Michael Hennebry | Wed, 18 Jan 2012 06:52:23 UTC

Heikki | Sun, 08 Nov 2009 11:16:54 UTC: "So what is the alternative to the DNS hack in the case of CDN?

HTTP redirects (including the already mentioned metalink/Http) simply do not work for everything because that is a per request thing. Let's say you have 100 1kB images on your web page. Were you using HTTP redirects, you would have to make 100 of them, each already as big as the resource it self. Not very efficient and adds a lot of latency to page loading." It seems to me that one would just need to redirect to the correct front door. Once one is at the correct front door, relative URLs would all go to the same site.


| Fri, 16 Jul 2010 05:11:45 UTC

The DNS is not a reverse Polish notation calculator. Or is it? http://bert.secret-wg.com/Tools/index.html

The DNS is whatever you want it to be, unless you're Paul Vixie. :)

Please remember, as bobel said, the DNS (using catchy names) has always been *optional*. (Internet) Numbers are the only thing that the network understands, and they are what we use when the use case is "mission-critical".

Apache, the world's most popular web server, has an option to disallow queries that use numbers. This is the ultimate "abuse" of the DNS: coercing people to use it.

Of course, it's easy to work around this silly option, using /etc/hosts.

Internet Numbers are not any more difficult to remember than phone numbers. The more you use them the more familiar and meaningful they become. In any event, hosts(5) provides for user-specified "unofficial names" and aliases. The DNS is not the only way to assign names to numbers. Hosts(5) can do it.

How many DNS queries will someone make in their lifetime? Imagine a number. Now imagine a file, a hash table, b-tree or some other data structure with a small entry for each of those lookups. How big would that file be? I'm quite confident it would be small enough to fit on your mobile phone, most likely, and certainly would fit on your netbook, laptop, desktop, workstation, etc. Maybe even entirely in RAM.


DNS Fan | Fri, 26 Mar 2010 02:57:06 UTC

There is an insightful and relevant blog entry (and clever graphic) that references this article at WhatTheHellSecurity.com titled "Security and the Unforeseen Use Case"


Eric Lawrence | Tue, 15 Dec 2009 00:32:55 UTC

Paul, if you have a network capture of the IE address bar behaving this way, I'd love to have a look (ericlaw AT microsoft).

Generally speaking, the IE address bar does not attempt to resolve against DNS as you type. The only exception is that, in IE8, after you've typed 4 characters, we'll attempt to resolve the hostnames for your top 4 previously navigated hostnames that match what you've typed thus far. We do not try to resolve hostnames you haven't visited, and we do not try to resolve partial hostnames based on what you've typed.

-Eric Lawrence, Program Manager, Internet Explorer Networking


Withheld | Mon, 30 Nov 2009 21:27:39 UTC

Why is the text gray? And why are half the comments gray text on a gray background? Why is the site designer trying to make the text hard to read?


Matt | Wed, 18 Nov 2009 17:14:52 UTC

Great stuff.


Paul Vixie (AUTHOR) | Wed, 18 Nov 2009 15:55:21 UTC

On CDNs:

I don't hate CDNs. I just don't think they need to hack in at the DNS layer. HTTP offers various kinds of redirects. IBM's WebSphere CDN takes this approach and has been a technical success. Anycast TCP is also in use by a few CDNs and works at least as well as any DNS-layer solution. Anycast is stable for minutes or hours at a time, so it's rare for two TCP packets to the same destination to reach different anycast contributors. My gripe about DNS-layer CDNs is that they push a lot of the service burden onto the clients and ISPs, and don't actually work better than available alternatives. Do 'dig www.microsoft.com' and count the number of times the zone changes inside the answer section. Every such change requires that a recursive nameserver somewhere restart its iteration. I realize that most successful businesses have to find ways to offload their costs, but this is ridiculous. How many CDN buyers actually compare the cost and benefit of CDN as compared to a single well provisioned well connected web server?


Paul Vixie (AUTHOR) | Tue, 17 Nov 2009 00:00:00 UTC

On CDNs:
I don't hate CDNs. I just don't think they need to hack in at the DNS layer. HTTP offers various kinds of redirects. IBM's WebSphere CDN takes this approach and has been a technical success. Anycast TCP is also in use by a few CDNs and works at least as well as any DNS-layer solution. Anycast is stable for minutes or hours at a time, so it's rare for two TCP packets to the same destination to reach different anycast contributors. My gripe about DNS-layer CDNs is that they push a lot of the service burden onto the clients and ISPs, and don't actually work better than available alternatives. Do 'dig www.microsoft.com' and count the number of times the zone changes inside the answer section. Every such change requires that a recursive nameserver somewhere restart its iteration. I realize that most successful businesses have to find ways to offload their costs, but this is ridiculous. How many CDN buyers actually compare the cost and benefit of CDN as compared to a single well provisioned well connected web server?


Paul Vixie (AUTHOR) | Tue, 17 Nov 2009 00:00:00 UTC

On RBLs:
I love RBLs. When Eric Ziegast invented the DNS RBL concept, it was the first wide scale use of DNS for something other than host names and host addresses. I'd like to see DNS carry many more kinds of information. But it would be a mistake to think that the policy data published in an RBL is the same kind of policy based response logic present in a CDN. Every DNS response we ever produced in the MAPS RBL was coherent -- if many people asked a question within a given TTL then they all got the same answer no matter where they were in the world or what we thought their connectivity was. I'm not against new kinds of data in DNS. I oppose incoherency, no matter how well engineered it may be.


Displaying 10 most recent comments. Read the full list here
Leave this field empty

Post a Comment:







© 2018 ACM, Inc. All Rights Reserved.