|WikiProject Computing / Networking|
I read the article and came away more confused and less informed than before. Also, there is no mention that the main strength of a dumb network is when the data transfer speed is “high enough” to handle any data traffic load, you do not need any intelligent prioritization of data. Everything gets through as soon as the data transmit it so the network can afford to be simple and “dumb.” This means less expensive equipment, no monitoring of data types, and no fancy communications protocols to follow (KISS). This is why many people see “dumb” as a smart choice for a network.
This article seems slanted against "dumb networks." It does not cite any references and the tone is unencyclopedic and confusing (see User:Septagram's comments, above). It has received only 14 edits (including my own) in the more than a year of its existence and the gross majority of the content was written by one anonymous user. I didn't nominate it for deletion, but were it nominated I would certainly vote to delete. Sean Hayford O'Leary 08:07, 16 December 2006 (UTC)
Text Removal by Originating Author
As the originating author, I have exercised my prerogative and removed my text from the site. The original article had been distorted so that it no longer reflected any insight into the properties of dumb networks and other network technologies
Deletion of Article
Apparently the dumb network is an article of faith for some people. The WIKI model fails when in the case of arrciles of faith since critical analysis cannot be performed on articles of faith.
This can be seen in the changes made by user Septagram which , though very earnest, are not accurate or complete. For example, SPAM is not currently dealt with on an end-to-end basis. Various technqiues, such as black holing, port disablement etc. are used to combat SPAM. All of these are examples of application-aware network techniques and are used to overcome deficiencies in the dumb nework mmodel. This article should be deleted beacuse it is inaccurate in its current state and cannot in any way be considered to be a reference article on dumb networks
For a more blunt assessment of the changes made by the use Sepatagram, he does not know the difference between a dumb network and the big pipe model of bandwidth allocation. —The preceding unsigned comment was added by 126.96.36.199 (talk) 15:36, 22 January 2007 (UTC).
Feel Free to Correct My Changes
I took the opportunity to review the past changes I made and do believe my changes where in earnest and where more an attempt to find balance, clarify and make it more readable. The intent of the article is to explain the concept of the Dumb network and SPAM is as you say, is an example of application-aware network techniques and are used to overcome deficiencies in the dumb network model. This fact should be included in the Dumb network article under the Con section for Dumb Networks, after the basic concept of a Dumb network is sufficiently explained. Concerning my knowledge, your comments seemed more of an AD Hominem attack and your efforts should be more directed towards correcting any “mistakes” I inadvertently made. Feel free to correct and expand the Dumb network article for people like myself to understand. Before I added my two bits, I noticed the articles’ concepts seemed unclear, slightly biased to Con, and rather lacking some in Pro. Please keep this in mind when correcting any “mistakes.” I believe it is important that this article is correct, because Dumb networks are referred several other articles and the obfuscation or deletion of this article would affect those several other articles. With the current charged events in network design, there may have been several attempts to slam some articles’ bias in one direction and any attempt to clarify the issues or find balance is squelched.Septagram 00:10, 23 January 2007 (UTC)
Updates of February 10th
I have tried to improve this entry by removing the references to bandwidth allocation as much as possible. The dumb network, contrary to the previous text, does contain multiple network protocols and these protocols are sensitive to the transmission and other needs of applications. TCP is tuned for error free reliable operation; UDP is tuned for minimum delay; RTP is tuned to minimize jitter and bandwidth for real time applications. Many 802 protocols at layer 2 operate for reliability, mobility, fairness etc. SIP at the session layer is being designed for collaborative applications.
I have tried to clean this article up as much as possible but, without a complete rewriting, it is not going to be an effective assessment of the capabilities and deficiencies of the dumb network model. The dumb network is a model of one method of application creation. It is not a model of network operation and it is especially not a model of bandwidth application. The dumb network is much more than a synonym for 'best effort' bandwidth allocation. —The preceding unsigned comment was added by 188.8.131.52 (talk) 18:59, 11 February 2007 (UTC).
Spam should be solved at the end points
Currently, spam can be solved at the endpoints. It even is, partly. Statistical filters are quite good, and computers could be secured against botnets, if only people knew how to use them (I'm confident they will, in a few decades). See the anti-spam techniques article for details.
The problem with centralized techniques is that false positives are effectively a limit to free speech. For instance, unconditionally blocking the outgoing 25 (smtp) port prevent people from sending e-mail by themselves. Of course they can (and do) ask an external service to forward their e-mail for them, but that is a potential privacy breach.