Talk:Software-defined networking

Portrait-in-gray (talk) 19:15, 1 November 2018 (UTC)

History Section

In light of this article [1], I think that SDN evolved out of both work from Stanford and much earlier work as well. I'd like to see that reflected in the history. Does anyone mind if I take a crack at that? — Preceding unsigned comment added by Feamster (talk • contribs) 18:25, 2 July 2014 (UTC)

That would be great, Nick. The modern emergence of the phrase SDN came from Kate Greene in this article [2]. In fact, she was the one that coined it. Your paper does a great job of outlining the intellectual history of work that lead to that, as well as was pulled in as the phrase started to become used more broadly. There is a very background of work that's completely ignored in this section. A synopsis of the origins paper would be fantastic. Martin casado (talk) 17:20, 28 November 2016 (UTC)

The history section is utterly and completely wrong. I strongly suggest letting Nick Feamster fix it, and let it reflect the true history of SDN best understood from his paper https://www.cs.princeton.edu/courses/archive/fall13/cos597E/papers/sdnhistory.pdf Sauravdas2 (talk) 05:22, 23 August 2015 (UTC)

Sounds good to me, go ahead and take an attempt at it. The history article right now is very poorly constructed. Shaded0 (talk) 18:19, 31 October 2015 (UTC)

The history has one huge problem, in that it is based on ignorance of other research and available standards. It lacks all reference to ITU-T standards that defines a complete SDN infrastructure used in ISDN, GSM and fibre optical transmission standards. On the project description page they claim that this is unique and no standards exists. KHF 18:54, 9 November 2015 (UTC) — Preceding unsigned comment added by Khflottorp (talk • contribs)

Unfortunately, this article *still* does not remotely reflect the history of SDN. Nick Feamster et al. article on history of SDN is a widely accepted account, used by almost any university or textbook covering SDN [3]. SDN became of widespread interest as a consequence of Martin Casado's PhD work, and the Ethane paper published at Sigcomm in 2007. The paper was recently recognized by ACM Sigcomm 2017 (the premiere networking conference) for a "test of time award" with the following citation: "Ethane ushered in the age of Software-Defined Networking (SDN) and a new generation of research that inspired both academia and industry to design network control planes that we can reason about." The term "SDN" was first used by Kate Greene in 2009 [4]. In its current form, the wikipedia page misrepresents the history and is confusing students of networking. In our classes, we have had to stop referencing Wikipedia because of the abundant errors in this page. Can someone please fix it? It's embarrassing in its current form... nickmckeown

In light of the above discussion, I am happy to volunteer to take up the pen to rewrite this section. Barring objection here, I will plan to replace the existing section with a shorter version of the history that's included in the paper that is referenced above. Feamster

Not sure about "separation of the control and data plane first used in the public switched telephone network as a way to simplify provisioning and management" in the first paragraph. Public telephone network has always separated them, right from when the control plane was girls pushing plugs in and out and data plane was analogue.JohnGrantNineTiles (talk) 14:12, 31 March 2018 (UTC)

Wow, if this is Nick Feamster, author of "The Road to SDN: An intellectual history...", I really like y'all's paper! I think that the perspective of SDN as one application of programmable network is important to the continuity of SDN research momentum, and including links to Active Networking will help, as you say, bring this research full circle. Please, crack away. :D

Revert 488130693

I have reverted the recent edits because they did not offer any improvements to the article. The lead section did not explain the concept in a satisfactory way, missed important aspects mentioned later on, and was partly unintelligible. The template boxes added, on general telecommunications, operating systems, and technology, did not offer a reader any advice for further reading, and were not directly related to the article content. If at all, a networking template could be added. Nageh (talk) 11:12, 19 April 2012 (UTC)

Yes, that is correct. Will try and work on this, but the exiting article is simply wrong — Preceding unsigned comment added by 87.68.250.67 (talk • contribs)

Thanks for giving it a try. Please keep in mind that Wikipedia does not permit original research. This means that it must possible to verify all statements to reliable sources. Also, all material must be written with a neutral point of view. Currently, the background section seems very opinionated, and still does little to explain the concept. Just consider the second sentence in the second paragraph, which reminds of marketing speech. Then there is a single sentence that attempts to explain the concept but doesn't, followed by a paragraph that assumes a reader to already understand it. If you have good sources on software defined networking, use them and cite them. Thanks again. Nageh (talk) 09:39, 13 May 2012 (UTC)

The definition is simple - decouple control from topology. The implications on designs currently on the table by vendors and usage by the various providers are not so simple. Needs more work. Need to adjust the tone in a few spots. No doubt.

Err, I may be a bit ignorant about computer networking, but the phrase "decouple control from topology" means precisely nothing to me. However I do know about the OSI reference model and I do know about TCP/IP, so I'm not completely clueless. Would it be possible to describe, in plain language, what is going on? Or at least to refer to other Wikipedia articles that explain the concepts in such a way that I can understand how they are being put together here? The current text reads a bit like an advertorial, or worse, a piece from a classical IBM manual from the "only clear if previously known" school of writing.

Given that software defined networking seems to have big practical implications (with the recent addition of Xsigo by Oracle), which were said by Open Source afficionadoes to allow Oracle to "exclude everybody else and leverage their massive installed user base.". Does this mean that SDN makes is possible to provide networking connections and authentification to end-users regardless of their IP address? If so I'd be happy if this was explained (in plain language). 82.92.220.29 (talk) 17:51, 4 August 2012 (UTC)

Need to add Wiki links as soon as language and tone are improved (in works)to: Internet Protocol (IP), Autonomous System (AS), Internet Engineering Task Force (IETF), Virtual LAN (VLAN), Virtual Private Network (VPN), Virtual Routing Forwarding (VRF), Open Flow Protocl (OFP), Directory Enables Networking (DEN), Virtual Machines (VM), iOS, Android, Hadoop, Map Reduce, Google, FaceBook, Internet of Things Architecture (IOTA), Managed Network Service (MNS), Infrastructure as a Service (IaaS)

Martin Casado

A possibly relevant article. Is Martin Casado really the inventor of OpenFlow / Software-defined networking? He isn't mentioned in either Wikipedia article. He also says: "I actually don't know what SDN means anymore, to be honest", i.e., it has become a meaningless marketing term over the years. Horatio (talk) 09:45, 30 April 2013 (UTC)

Yes, good reference to add. It does appear to come out of a student project, and then turned into hype by the Silicon Valley Vulture Capitalists? Needs work certainly. W Nowicki (talk) 23:01, 21 May 2013 (UTC)

Yes, I wrote the original version of OpenFlow, and the predecessor which was used in my PhD research project called Ethane. And the emergence on SDN in the current context was coined by Kate Greene in this article [5] which was based on that work. However, as Jennifer Wexford's and Nick Feamster's paper correctly illustrates, there was a very rich intellectual history that both lead to my work, and was pulled in independently under the SDN umbrella once the term was popularized. I think a fair coverage of the origins would acknowledge that software based networking has been around since networking has, and that there have been many architectures that have included aspects of modern SDN design, however the recent emergence came out of a long line of academic work.Martin casado (talk) 17:19, 28 November 2016 (UTC)

NPOV Dispute

This article appears to be mostly lifted from this industry paper: https://www.opennetworking.org/images/stories/downloads/sdn-resources/white-papers/wp-sdn-newnorm.pdf.

The paper itself proclaims that it is authored by an industry group called the Open Networking Foundation. Moreover, the paper (and consequently the Wikipedia article) reads like marketing copy. I came to this article after seeing the term "software-defined network" in a news story. After reading the article, I am still confused about the meaning of the term. While I am unable to contribute to the improvement of this page, I hope that someone else who is more knowledgeable on the topic can do that.

Wajlee (talk) 23:41, 13 August 2014 (UTC)

SDN has become a term that is generally accepted by the IETF as new thought on the programmable networks paradigm. The thought on this is quite far reaching beyond commercial interests and really needs to be clarified for the layman instead of rejected by the people who find issue with copy. The fact is, many scientific topics and the terminology behind concepts are first brought to light for the interested pubic by folks with a mind for marketing who are not always tactful but are often good meaning and at least contribute. Most who read this will want a top level or partial understanding of the science and truly desire to know the potential improvements to their lives and the status quo that the science can offer. (In this case there is much it can offer) It is this knowledge that aids in the understanding of the concepts for both the scientist and those interested. The article does need clarification from people who are both unbiased and knowledgeable but I'm glad this article exists I hope it continues to be improved upon.

The following reference is probably the most authoritative and unbiased on the subject and in my opinion, provides a valid platform for the legitimacy of the term SDN. (Coming right from the IETF) The great people who wrote it should really should be pinged for this Wiki because they are the most qualified to comment on the science behind the term and enlighten all of us.

http://www.ietf.org/id/draft-haleplidis-sdnrg-layer-terminology-07.txt

There really is allot to add to the Wiki such as new thought on "Recursive Network Architectures" which are really exciting. I will agree that after reading this, it feels quite incomplete but while I am perhaps more knowledgeable than most, surely not enough so to do the topic justice.

Zhabrat (talk) 19:42, 2 October 2014 (UTC)ZACHZhabrat (talk) 19:42, 2 October 2014 (UTC)

Totally agreed, the article needs a detailed rework so it's less of "SDN is the best thing since sliced bread" story. It surely is a great thing, and the article isn't that bad, but the content and style should be improved. — Dsimic (talk | contribs) 22:50, 26 October 2014 (UTC)

The "Limitations of current networking technologies" section

This whole section (which makes up about one third of the article) is a verbatim copy/paste from the ONF "Software-Defined Networking: The New Norm for Networks" whitepaper, which is listed as reference number 1 (this one here). I strongly suggest that this be removed completely. --Darkstar (talk) 13:49, 15 May 2015 (UTC)

To hyphenate or not

Anyone else find it interesting... there are 2 spellings of this, commonly used? This article has 9 external links with SDN spelled out - basically half with hyphens, half without. Is it random that the article commits to the hyphenated one? It seems intentional (I see the same on "Software-defined Protection." Is there a grammar policy page regarding this? It seems like terms (capitalized in sentences, and have popular acronyms?) like this are something like proper nouns, falling outside the simple in-sentence grammar rules. Thoughts? Thanks. Justapersona (talk) 20:53, 21 May 2015 (UTC)

Hmm. I guess, first of all, this isn't a proper noun. I wonder what it is called, when people capitalize things that are not proper nouns (like people have sometimes, with this term). I love this definition of Proper_noun, "A proper noun is a noun that in its primary application refers to a unique entity, such as London, Jupiter, Sarah, or Microsoft, as distinguished from a common noun, which usually refers to a class of entities (city, planet, person, corporation), or non-unique instances of a specific class (a city, another planet, these persons, our corporation)." Justapersona (talk) 21:10, 21 May 2015 (UTC)

Hello! It should be hyphenated, and not capitalized – "software" modifies "defined", and they both (as a compound modifier) modify "networking". Please see MOS:HYPHEN for more details on hyphenation. — Dsimic (talk | contribs) 21:15, 21 May 2015 (UTC)

Concept

The text is identically with the one from https://www.opennetworking.org/sdn-resources/sdn-definition. That should not be the case. — Preceding unsigned comment added by 87.178.176.99 (talk) 15:11, 13 June 2015 (UTC)

The text was apparently modified in the meantime, but everything beyond "dynamic" and "manageable" in the first sentence is still complete marketing drivel. SDN is neither more cost-effective, nor more suitable for high-bandwidth. And I don't even know what "dynamic nature of today's applications" is supposed to mean. I suggest to change the first sentence to "SDN is an architecture purporting to be dynamic and manageable." Marcello Caruso (talk) 17:30, 13 December 2016 (UTC)

"The New IP"

The reference to "The New IP" (a marketing term) has been deleted. This looks like a paid edit: https://en.wikipedia.org/wiki/Special:Contributions/67.82.50.114 Nickhilliard (talk) 15:37, 3 September 2015 (UTC)

Introduction - first sentence

With regard to the sentence below:

"Software-defined networking (SDN) is an approach to computer networking that allows network administrators to manage network services through abstraction of higher-level functionality."

I understand abstraction to be a detachment from lower-level functionality to provide higher-level functionality. For example, the IDE interface of a hard disk drive is an abstraction of lower-level functionality implemented in the disk controller (seeking, positioning, reading and writing) onto a higher-level of functionality, consisting of a single-dimension array addressable through a logical block address.

I would re-write the first sentence as follows:

"Software-defined networking (SDN) is an approach to computer networking that allows network administrators to manage network services through abstraction of lower-level functionality."

Edepa (talk) 22:22, 18 February 2016 (UTC)

Agreed - I'll make the change.Timtempleton (talk) 17:58, 22 June 2016 (UTC)

I think that first sentence should describe what software-define networking is at its core, no?

"Software-defined networking is a computer networking approach that separates the routing logic, handled by the control plane, from data movement, managed by the data plane. In this model, software governs both aspects, enabling programmable and dynamic network management."

SDN is fundamentally about decoupling the control and data planes, allowing for programmable and dynamic network management. This can be as complex as managing global cloud infrastructure or as straightforward as a software-implemented Ethernet bridge in the Linux kernel.

The existing first sentence suggests that people use SDN "in order to improve network performance and monitoring, in a manner more akin to cloud computing than to traditional network management." But that's just one common use-case. Another common use case, is to use a software defined network to simplify dev ops. For example, loads of people let docker-compose create a network for their application. All of these things are enabled by software defined networking. Cheese53 (talk) 07:39, 13 September 2023 (UTC)

The existing first sentence is full of jargon. Control vs. data plane is described in the second sentence but there is also clouded by jargon. The suggestion by @Cheese53 would be an improvement. One complication is the rest of the article is written in the same indirect and jargony style as the lead so the current lead is arguably a good summary of the article. ~Kvng (talk) 14:36, 17 September 2023 (UTC)

Introduction - Second Paragraph

I extended the previous version, which was:

SDN requires some method for the control plane to communicate with the data plane. One such mechanism, OpenFlow, is often misunderstood to be equivalent to SDN, but other mechanisms could also fit into the concept.

to:

SDN was commonly associated with the OpenFlow protocol (for remote communication with network plane elements) since the latter's emergence in 2011. Since 2012[1][2], however, many companies have moved away from OpenFlow as a single-solution, and have embraced a number of different techniques. These include Cisco's Open Network Environment and Nicira’s Network Virtualization Platform.

My rationale is that:

• SDN does not limit itself to control + data planes any more
• Since OpenFlow was mentioned, and is a well known association, a substitute or cross-reference to current-trends may be useful.

I ought to have posted here before changing, sorry about that. Do give your feedback and I'll factor them in.

Obscure-helper (talk) 16:19, 7 March 2016 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified one external link on Software-defined networking. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Corrected formatting/usage for http://www.websprocket.com/installation_guide.doc

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—^{cyberbot II}_{Talk to my owner:Online} 10:29, 2 April 2016 (UTC)

"The need for a new network architecture" is copied word-for-word from a whitepaper

And is not quoted:

See, "A need for a new network architecture" at https://www.opennetworking.org/images/stories/downloads/sdn-resources/white-papers/wp-sdn-newnorm.pdf — Preceding unsigned comment added by 65.112.8.140 (talk) 20:49, 28 July 2016 (UTC)

Introduction - First Paragraph

It should be mentioned early in this article that SDN causes a paradigm shift in the Ethernet and the IP Layer, from distributed control (where network elements "talk" to each other using some protocol like OSPF, IS-IS, STP, etc., and then each element configures itself based on the gathered information) to centralized control (where every network element is configured by a central intelligence, that keeps track of the entire network status). This is in my opinion the most central point to understand SDN, as it actually substitutes distributed protocols. In other words, the intelligence is taken out of the individual nodes and put into a central network master. Abstraction (which in the article currently appears to be the most important point of SDN) is in my opinion at least a questionable characteristic. What exactly is the abstraction provided by SDN? Applications that run on top of the central controller still work on the level of flows and topology, very much like the controller in each OSPF or IS-IS node. Marcello Caruso (talk) 17:30, 13 December 2016 (UTC)

regarding the intro, is SDN is an approach to "cloud computing"? isn't SDN rather an approach to networking? this part seems disturbingly out of place. Portrait-in-gray (talk) 19:16, 1 November 2018 (UTC)

I agree to the previous statements in this talk page paragraph, i.e. shift from distributed to centralised control etc., and the "cloud computing" remark is either misleading or requires explanation/qualification/deletion. Some mention of network disaggregation, DevOps, network automation and network programmability could also be made, as some believe those are the key concepts to SDN. Veterans of telecoms may recall the Intelligent Network architecture which seems analogous to me, where the "paradigm shift" was moving the control functions from the switches (SSF/SSP) to "general-purpose" computers (SCF/SCP) by using "Application Part" signalling protocols such as INAP. Perhaps a more enlightened editor/author could make the necessary changes to this SDN wikipedia article? Thanks in advance. 82.21.133.132 (talk) —Preceding undated comment added 21:57, 2 December 2018 (UTC)

NETCONF

Should there be some reference to NETCONF? Is OpenFlow the only SDN protocol? 167.98.51.116 (talk) —Preceding undated comment added 17:13, 30 October 2019 (UTC)