Happy New Year, everyone.  Only six more years till hoverboards! 

At the beginning of the new year, there’s almost always a slew of tech news articles and editorial talking about how 2009 will be the year of X. 

To wit, it has been the year of “the Linux Desktop” every year since 1999.  (Personally, I think that 2008 was the year of the Linux Desktop, but that’s an article for a different day.)

But while predictions may provide some source of humor when they go awry, the paradox is that they should be taken seriously.

Part of the problem with making predictions is that if the result is positive, they can become self-fulfilling, negative, and they become self-avoiding.  Look at Y2K.  The world thinks nothing happened – however, if you were in IT during the late 1990s, you know that Y2K required a major overhaul, increased upgrades, etc.  Y2K wasn’t a disaster overhyped, it was a disaster avoided through massive hard work.  It also may have had no small part in the zenith of the tech boom of the 1990s, as companies bought new hardware before the usual refresh cycles. 

Network World is making one big prediction that I think does deserve some attention. 

The first is Steve Taylor and Larry Hettick’s prediction that 2009 will be the year of IP video. 

Our observations: We recall the days when public and enterprise networks were engineered first for voice and data second, but as data traffic demands grew, the engineering focus by necessity had to change to data first, voice second. We see the same evolution in network engineering focus as video demand grows to surpass data. We also note that while some enterprise IPV will be sent and received entirely across private networks (especially for telepresence), inter-company and business-to-consumer traffic will principally cross the Internet - so service providers will need to accommodate both consumer and enterprise video traffic in a way that does not compromise voice and data network integrity. And with the consumer market for Internet delivered commercial video also beginning to burgeon, the task of managing all the video traffic across the network cores will not be trivial.

I suppose that now would be a good time to mention that we’re probably going to be stepping up our Whiteboard Video series articles in the new year. 

In addition to teleconferencing, video’s simply a simple way to convey information to those who learn by both sight and sound.  It’s much easier to explain a concept when you can show them, rather than just telling them about it.  Maybe one new years resolution could be to check with your marketing and sales departments to see if they have any plans to put video online for customers, or with HR to see if there’s any important training videos going out on the intranet. 

The problem is that video, voice, and data all travel on the same pipe.  If one of those three monopolizes the pipe at the expense of the other two, it doesn’t matter how large or how small the pipe is.  Network monitoring and proactive management is sorely needed in a “triple pipe” environment.  Proper network management allows you to mitigate the worst problems with an oversubscribed line until you can get more bandwidth; but improper network management will cause problems no matter how much bandwidth you add. 

by Brian Boyko

Several people have suggested that I take a look at the Wall Street Journal story, or a number of stories based on that story, about how Google is allegedly abandoning network neutrality in favor of a “fast lane” of Internet traffic. Specifically, John C. Dvorak over at Network World asked “Why has Google demanded a ‘fast lane?’” It seems to be based on reporting by Vishesh Kumar and Christopher Rhoads over at the Wall Street Journal. The opening paragraph of that story:

The celebrated openness of the Internet -- network providers are not supposed to give preferential treatment to any traffic -- is quietly losing powerful defenders.

I was once the associate editor of the Daily Texan, a daily newspaper in Austin, Texas. In the Newspaper business, we call the opening paragraph the “lede,” and in the lede, you are supposed to inform the reader of the most important aspects of the story.

Kumar and Rhoads got the lede very wrong. Here’s other important points from the article.

Google Inc. has approached major cable and phone companies that carry Internet traffic with a proposal to create a fast lane for its own content, according to documents reviewed by The Wall Street Journal. Google has traditionally been one of the loudest advocates of equal network access for all content providers.

At risk is a principle known as network neutrality: Cable and phone companies that operate the data pipelines are supposed to treat all traffic the same -- nobody is supposed to jump the line.

Also from the article:

In addition, prominent Internet scholars, some of whom have advised President-elect Barack Obama on technology issues, have softened their views on the subject…. But Lawrence Lessig, an Internet law professor at Stanford University and an influential proponent of network neutrality, recently shifted gears by saying at a conference that content providers should be able to pay for faster service. Mr. Lessig, who has known President-elect Barack Obama since their days teaching law at the University of Chicago, has been mentioned as a candidate to head the Federal Communications Commission, which regulates the telecommunications industry.

And arguing that Google wants to violate the spirit of network neutrality:

Google's proposed arrangement with network providers, internally called OpenEdge, would place Google servers directly within the network of the service providers, according to documents reviewed by the Journal. The setup would accelerate Google's service for users. Google has asked the providers it has approached not to talk about the idea, according to people familiar with the plans.

Asked about OpenEdge, Google said only that other companies such as Yahoo and Microsoft could strike similar deals if they desired. But Google's move, if successful, would give it an advantage available to very few.

Now, here’s what’s wrong with each of these points.

Lawrence Lessig referred to the story as “the made-up dramas of the Wall Street Journal.” (“Made-up” is never a phrase you want to associate with journalism.) In his blog, he points out:

The article is an indirect effort to gin up a drama about a drama about an alleged shift in Obama's policies about network neutrality. What's the evidence for the shift? That Google allegedly is negotiating for faster service on some network pipes. And that "prominent Internet scholars, some of whom have advised President-elect Barack Obama on technology issues, have softened their views on the subject."

Who are these "Internet scholars"? Me. And of course, because I have "softened" my views about network neutrality, and because I advised the Obama campaign about technology issues during the primary, it follows (and obviously so) that Obama too must be going soft on network neutrality….

But the whole punch of the story comes from the suggestion that my position is something new. … Missing from the article, however, is the evidence that my view is a "shift" or "soften[ing]" of earlier views. That's because there isn't any such evidence. My view is the view I have always had -- whether or not it is the view of others in this debate.

Eek. Strike one. What about Google? Well, in Google’s Public Policy Blog, they explain:

Despite the hyperbolic tone and confused claims in Monday's Journal story, I want to be perfectly clear about one thing: Google remains strongly committed to the principle of net neutrality, and we will continue to work with policymakers in the years ahead to keep the Internet free and open.
P.S.: The Journal story also quoted me as characterizing President-elect Obama's net neutrality policies as "much less specific than they were before." For what it's worth, I don't recall making such a comment, and it seems especially odd given that President-elect Obama's supportive stance on network neutrality hasn't changed at all.

Here’s one of the areas where the Wall Street Journal makes it’s mistakes: Google’s current “OpenEdge” project is basically a form of “edge caching” – that is, providing a copy of their offerings that is physically closer to the end-user, by hosting it on broadband provider’s own local networks. Yes, this would create faster access to Google, than to it’s competitors that do not edge-cache but this is not a violation of network neutrality.

The principle of network neutrality is not violated when a company invests money to improve the performance of the services they offer over the open Internet. Where Network Neutrality comes into play is when a company pays money (or a broadband provider demands money from a company) in order to provide added performance of services at the expense of their competitors.

By moving the data physically closer to the end-user on the network, Google lowers the propagation delay of their offerings. In fact, moving data physically closer to your users is the only way to lower propagation delay, at least until such time as we can figure out how to move data faster than the speed of light.

But in so doing, Google does not slow down the offerings of Yahoo or Microsoft. Edge-caching is not a zero-sum game; if it took 100ms to reach Google, Yahoo, and Microsoft earlier, it will still take 100 ms to reach Yahoo and Microsoft even if it only takes 25ms to reach Google’s cached servers. This is a good thing – a new technology which can improve the end-user experience.

And we want these kind of network improvements. If Google does, someday, indeed invent warp-speed data transmission, it would be hard to argue that just because Google takes less time to access than competitors that it is somehow violating network neutrality. Violating laws of physics, maybe, but not network neutrality.

What would be a violation of network neutrality is if, somehow, Yahoo and Microsoft were to suffer extra delay due to the improvements granted to Google. Using the above example, where Google, Yahoo, and Microsoft all have 100ms of latency to the end user, when Google uses something like packet shaping or QoS policies based on source (Layer 8?) to decrease the latency of their own packets – say to 75ms - in order to avoid congestion, Yahoo and Microsoft necessarily have their own packets delayed a bit, say to 105ms. That scenario is a zero-sum game.

Now, there are some scenarios where QoS prioritizing and packet shaping might make sense. For example, VoIP, teleconferencing, and gaming over the internet all require low-latency, low-jitter connections and performs best when using smaller packets with little delay. An ISP could put these latency sensitive applications into a separate class of traffic, and in so doing, degrade the response time of latency non-sensitive applications.

Does this violate network neutrality? No – because the preferential treatment is being assigned equally to all applications – all competitors – who design these latency sensitive applications. World of Warcraft may get lower latency under this scenario than Google Docs, but it does not get lower latency than Eve Online, in other words. Or put another way, if you simply give a higher priority of traffic to UDP applications over TCP applications, you’re making decisions based on OSI layer 4, not OSI layer 7. The other thing about this setup is that the prioritization of the traffic is based on the needs of the network application, not on the ability for the application provider to pay. World of Warcraft would make better use of lower-latency connections than Google Docs would.  (I know, I know, it sounds vaguely Marxist, but it’s still a good idea.)

It’s complicated, thorny, and hairy. However, I’m going to propose a rule of thumb. Call it Boyko’s Network Neutrality Guideline if you like, because there aren’t enough things in this world named after me.

“A particular method of improving Internet performance for a particular application violates Network Neutrality only when the implementation of the method would degrade the performance of competing applications."

Royalty checks to me can be sent c/o Network Performance Daily.

This article includes research contributions from Joel Trammell.

The Obama-Biden transition team promised last Monday, Dec. 8^th, that they would provide most policy documents from meetings with outside groups – i.e., lobbyists – would be posted on the Change.gov Web site.

By Wednesday, Dec. 10^th, this policy already saw some interesting results. David Kravets over at Wired’s Threat Level blog pointed out that the site has already published a paper detailing the requests of the MPAA’s lobbying organization, which include requesting filtering information from technology companies.

We’re not against the MPAA using the means available to protect their intellectual property concerns, but there are two problems with filtering: false positives, and performance degradation.

False positives are already a major problem with the content industry – back in 2003, the RIAA sent a cease and desist letter to Penn State University – they had confused work from Prof. Peter Usher at the Department of Astronomy and Astrophysics with that of Usher, the R&B pop singer.

This is also a recent problem; in October of 2007, Google launched a copyright filter for the YouTube Web site. It, too, has many false positives. For example, a fan production of the reality TV show “The Mole” was removed, presumably, because it was confused with the real thing by the filter. Judging from the production values of the fan-film, it’s very unlikely that a human censor would confuse the two.

(Fun fact I learned while researching this article: Andy Warhol made a “Batman” fan film back in 1964.)

Videos removed for copyright complaint – legitimately or not - have been catalogued (but not archived) at YouTomb, a project from MIT Free Culture.

But YouTube is one, privately operated Web site. Filtering the content as it is uploaded merely affects the time to publish, not the time to distribute. Additionally, videos can also be hosted on competing sites.

If one were to try to use filtering on the Internet as a whole, as the MPAA seems to be lobbying, it is likely that the results would be similar to the results of the tests run by the Australian government – where even the best of filters degraded network performance, and the better the filter was at avoiding false positives and false negatives, the more performance degraded. Even the best filter wasn’t very effective.

The lesson to learn from all of this is that too often, measures taken in the name of “computer security” – even if it’s to instill a false sense of security – can have serious impacts on network performance. For this reason, those in the enterprise responsible for making sure that networks remain secure and those responsible for making sure that applications remain responsive absolutely need to coordinate efforts.

Thomas Nolle at ComputerWorld (via NetworkWorld) suggests that not only will the economic downturn affect IT budgeting, but that networking, in particular, will take a harder hit than the rest of IT.

The logic goes something like this: when the first tech bubble burst in the early 2000s, IT spending shifted from networking to computer systems and software. To quote Nolle:

The fact that the point where the shift occurs corresponds with the previous major economic downturn raises some legitimate questions about whether networking might not take a further hit in the current slump, as well as questions of what might be done to prevent that.

That is one theory.

But I think it’s simply more likely that there were other factors that precipitated 2000s IT spend shifting to computer and software expenditures.

For example, 1999 and 2000 were the years of the Y2K scare. (I still believe that Y2K will, of course, eventually kill us all. It just didn’t happen on New Year’s Eve, 1999, because everyone knows that Y2K will strike when you least expect it...)

To prepare for Y2K, companies spent millions on upgrading their entire IT departments to newer equipment that was “Y2K compliant.” It makes a bit of sense that more was spent on the desktop than in the network – there’s only a handful of data centers but tons of workstations.

Additionally, Windows 2000 came out in February of 2000, with Windows XP soon after in October 2001. Both OSes provided a more stable, and thus, more business-friendly computer working environment – so companies might have a compelling reason to upgrade.

Or, consider that prices for desktop computer hardware, already on a deep decline, started hitting very low prices, comparatively, around the same time – computers were becoming so cheap that there were companies that would give you a computer with 2 years subscription to an Internet service. At those prices, computers could be given to every employee instead of only the most savvy. Also, 2000 was when early graduates of universities in the Internet era were out looking for jobs – and these graduates knew how to use PCs, which justified the cost.

So I think that perhaps Nolle might be confusing correlation and causation. Then again, Nolle may be right and I might be confusing correlation and causation. Then again, correlation and causation might be causing confusion. (Then again…)

Additionally, the networking environment of 2000 is very different from 2009. How many applications did your company have on the network in 1999/2000? How many does it have today? Can you even count that high? 2000 was before the advent of Salesforce.com and other SAAS products that depend on network connectivity – back then, you were just as likely to e-mail a file as you were to copy it to a floppy disc.

And let’s not forget the point about the number of people working remotely, which will actually be more important as companies shrink campuses.

However, that doesn’t mean he’s necessarily wrong about some of the points later on in the article. For example:

The question we might ask is why networking couldn't capitalize on the attention it received. The answer, I think, lies in the stuff that binds networks to applications. The pivotal point in that critical issue came in the early 1990s, when IBM's Systems Network Architecture was supplanted by TCP/IP. SNA network equipment was just too expensive, and enterprises went to the lower cost of TCP/IP instead. The critical thing was that SNA was an application architecture as well as a network architecture, and TCP/IP vendors didn't present application tools… Networking won hearts and minds in the '90s, then lost them again because it didn't offer the whole solution. The application connection to the network was never made by the network vendors, and so IBM and other system and software players continued to control that critical linkage -- and still do today.

We often say, (mostly because we agreed with it when Jim Metzler said it,) that in IT, you either develop applications or you deliver applications. It’s all about the applications – because ultimately, layers 1-6 have no purpose unless they’re supporting layer 7.

If you’re going to have problems with IT budgets during the economic downturn, the best way to weather the storm is to make it clear how the network enables the applications that run on it, and how the applications add to the business’s bottom line.

If anything would precipitate a slowdown, it would probably be that for years, during the good times, CIOs have been future-proofing their networks in order to meet increased demand during a time when they couldn’t just throw more resources at the problem – that day seems to have arrived, so now they may be looking to finally use the capabilities that they paid for when times were tougher.

Thanks to Chandra Hosek and Steve Harriman for their help in writing this article.

Readers of this blog will remember “The Strange Case of Julie Amero,” which we’ve covered extensively here:

• Introduction
  
• Commentary by defense expert Herb Horner
  
• Quick info from CEO of forensic software
  
• Commentary by Prosecution Expert Det. Mark Lounsbury
  
• Conclusions

Julie Amero’s conviction was overturned after the Internet community, led by Alex Eckleberry of Sunbelt Software, rallied around her cause.  It’s rare for a judge to throw out a case after a conviction, but the evidence was overwhelming.  A new trial was ordered.

You could have argued that the prosecutors in this case were computer illiterate, but for months, prosecutors held the threat of a new trial over Ms. Amero’s head, and then, instead of dropping the charges, to suggest a plea bargain - $100 fine, loss of state teaching license, and a conviction for disorderly conduct. 

Ms. Amero took the plea bargain, and with the case concluded, she finally was able to speak out in an interview with ComputerWorld.  Here’s a bit of information that we didn’t know:

What was on the screen?
Little itty bitty tiny pictures of sites: Viagra sites, sex enhancement creams, women in lingerie, things of that sort. Nothing lewd.
So no pornography?
No.
Was there nudity?
There was no nudity. There were sites listed. And the things they said [in court] I clicked on and went and looked at have been proven that they never were clicked on and looked at. The things that were on there were just inappropriate things to be looked at in a classroom; Victoria's Secret kind of stuff, you know….

So there was never anything pornographic?
[The prosecution] said there was one site visited, where there was a thumb-sized picture of oral sex.
So they found one picture of oral sex on the computer, but you didn't see that?
No.

The prosecution in this case knew full well that Ms. Amero was completely innocent.  And had an opportunity to try to mitigate the damage by dropping the charges when the wrongful conviction was overturned after the evidence came to light.  They did not.  And eventually they got what they wanted – some sort of conviction of an innocent woman for a crime that turned out never to have happened.

From Rick Green at the Hartford Courant:

New London County State's Attorney Michael Regan told me late Friday the state remained convinced Amero was guilty and was prepared to again go to trial.

"I have no regrets. Things took a course that was unplanned. Unfortunately the computer wasn't examined properly by the Norwich police," Regan said.

"For some reason this case caught the media's attention,'' Regan said.

The good news is that though it may not have resolved satisfactorily, at least it is finally resolved.

President-Elect Barack Obama, recently put a new video on Change.gov, the official Web site of the office of the President-Elect. In the video, Obama is seated in the office of the President-Elect, sitting in the chair of the President-Elect in front of the desk of the President-Elect. And if I had to guess, he’s probably reading prepared notes from the teleprompter of the President-Elect into the YouTube camera of the President-Elect.

These YouTube videos aspire to function much like an online, 21^st century version of the FDR’s “Fireside Chats.” Coincidentally – or perhaps not - Roosevelt’s first Fireside chat, broadcast in 1933, was entitled, On the Bank Crisis. This was also the subject of Obama’s broadcast. And like FDR, Obama is proposing a federal employment program much like Roosevelt’s New Deal.

During the New Deal, the Civilian Conversation Corps, or CCC, was a work-relief program designed to keep people employed with a semi-steady paycheck. One of the main ways that the CCC employed young Americans was by putting them to work solving the environmental problems of that era – which mainly involved flood prevention, soil erosion prevention, wildland fire suppression, reforestation, et al. You can go to Wikipedia for the full details.

Similarly, Obama’s plan, (or the sketches of it outlined in the short YouTube video), calls for increased energy efficiency in national infrastructure. Obama didn’t mention it specifically, but it’s not a far guess to think that part of this process will be “greening” Federal IT – and that usually means server consolidation (perhaps aided with WAN Optimization) and virtualization – to do more with less of a power draw – as well as putting the network to new uses, such as teleconferencing instead of spending money on airfare. Both of these will require network monitoring and adept management.

What Obama did specifically mention:

“It is unacceptable that the United States ranks 15^th in the world in broadband adoption.”

Hmm. Sounds familiar. Sounds like something we’d say. (Mr. President-Elect, are you subscribed to our RSS feed?)

The CCC, and similar program WPA, also focused on building and improving the infrastructure of the country – broadband improvements, of course, are improving digital infrastructure. (Of note – American broadband improvements must be done on American soil, and can therefore be almost “outsource-proof.”)

Also of interest was Obama’s pledge that:

“In addition to connecting our libraries and schools to the internet, we must also ensure that our hospitals are connected to each other through the internet. That is why the economic recovery plan I’m proposing will help modernize our health care system – and that won’t just save jobs, it will save lives. We will make sure that every doctor’s office and hospital in this country is using cutting edge technology and electronic medical records so that we can cut red tape, prevent medical mistakes, and help save billions of dollars each year.”

We at NetQoS have a little bit of familiarity with how difficult medical networking needs can be. In Volume 3 of Performance Edge Journal, we published a case study of OSF Healthcare, which is a network of multiple acute care, long-term care, and college of nursing facilities and also a primary care physician network.

One of the applications that OSF wanted to implement included Picture Archival and Communications System (PACS) – in order to send and manage very large cardiac images. The application had very slow response times, and you know, when something’s wrong with your ticker, time’s of the essence.

Anyway, NetQoS considers this one of our big success stories. Through SuperAgent, they found that the delay was caused by excessive server response times, and using ReporterAnalyzer, they were able to figure out that some cardiac images were being sent to different sites across the WAN and not stored locally, slowing down retrieval times considerably, taking up large amounts of bandwidth unnecessarily.

Whether or not Obama’s plan will actually work, at least he seems to be aware of the real need for network performance in the healthcare industry.

At the risk of waxing political; Obama talking about technology policy means that the next four years (at least) will be interesting times for geeks. Under the Clinton and Bush administrations, discussions of what should be done about networking issues were mostly confined to – well, to blogs like this. (And even then, it was mostly confined to Slashdot). But, whether or not you agree with him, Obama seems to be using the power of his office – or of his future office – to call attention to the problems that up to now, only us techies were paying attention to.

Update: Texas PI Law

Benjamin Wright, (an advisor to an electronic-discovery firm Messaging Architects) posted a comment to our coverage of the Texas law that requires companies which “investigate” computers (which could possibly mean PC repairmen, although that wasn’t the original intent of the law) to have investigative licenses – the “Texas PI law” for short.

He pointed out that already there is one unintended consequence – and that is that those caught in red-light traffic cameras are suing the manufacturers because even though the red light footage is “evidence of a crime,” some (most?) traffic-cam organizations are not licensed to act as a private investigator, and therefore it is illegal for them to present that computerized evidence in court.

When the Texas PI law first came out, we were concerned about “unintended consequences,” and those consequences striking computer repairmen, network engineers, sysadmins, and others. This, on the other hand, seems to be the intended consequence; the red-light camera company was doing forensic work to be used in court dealing with computers. (Let’s ignore the merits and drawbacks of red-light cameras for right now.)

It is, however interesting to note that the company named in one fine appeal, American Traffic Solutions, is based in Arizona, which explains, partially, why they may have not gotten a PI license in Texas.

The American Bar Association has weighed in on this issue [PDF], arguing that computer forensics experts should not need PI licenses for forensic work.

Update: Respecting the Network Engineer

Yesterday, we published Chandra Hosek’s column on how Network Engineers often get less respect than they deserve. Thomas Nolle at Network World has published a column suggesting some of the reasons why.

Nolle notes that since the first Tech bubble burst in 2000, computer systems and software have gotten larger than average shares of investment, while networks have gone down since then. From the article:

The question we might ask is why networking couldn't capitalize on the attention it received. The answer, I think, lies in the stuff that binds networks to applications. The pivotal point in that critical issue came in the early 1990s, when IBM's Systems Network Architecture was supplanted by TCP/IP. SNA network equipment was just too expensive, and enterprises went to the lower cost of TCP/IP instead. The critical thing was that SNA was an application architecture as well as a network architecture, and TCP/IP vendors didn't present application tools… Networking won hearts and minds in the '90s, then lost them again because it didn't offer the whole solution. The application connection to the network was never made by the network vendors, and so IBM and other system and software players continued to control that critical linkage -- and still do today.

This brings new perspective on the idea that, as Jim Metzler put it, in IT you either develop applications or deliver applications.

Update: Australian Internet Fliters

Computerworld reports that the Australian-based Electronic Freedom Project is organizing protests on the 13^th of December to protest Internet content filters mandated by the government. At the same time, the Minister for Broadband, Communications, and the Digital Economy, Sen. Stephen Conroy, has given his explanation of the ISP filtering plan; but ultimately, “Teh Moges” at Slashdot pretty much sums it up with: “Any efficient filter won’t be effective, and any effective filter won’t be efficient.”

Sen. Conroy earlier tried to place pressure on Mark Newton’s employer to keep the Australian network engineer’s criticism quiet.

Update: Groups push for Net Neutrality legislation in Obama’s administration

During the 2008 presidential campaign, then candidate Barack Obama pledged support for Network Neutrality principles and regulations. Now that he’s President-Elect, the Open Internet Coalition is calling on Obama to follow through, according to Network World.

A column in The Register claims, amongst much wailing and gnashing of teeth, that implementation of BitTorrent-over-UDP (dubbed uTP) in the new alpha version of uTorrent, one of the official BitTorrent client applications, will end the Internet as we know it and completely congest the network.  The title is “Bittorrent declares war on VoIP, gamers.”

Considering the fact that BitTorrent, Inc., has, if anything, always gone out of it’s way to avoid declaring war on anybody, this seemed to me a little bit odd.

I’ll admit that it kind of worried me – TCP has traffic congestion management built into the protocol, UDP does not.  When UDP and TCP exist on the same network (for example, when rolling out VoIP on a corporate network), QoS policies are needed to keep UDP from taking up all the bandwidth while TCP meekly  throttles back.   Jim McQuaid has a Whiteboard series video up about it, and called it “Nice Guys Finish Last.”

The reason that UDP is popular is that it’s a lightweight protocol that doesn’t do much handshaking.  It sends the data “that-a-way” and doesn’t particularly care if it makes it.  That makes it perfect for VoIP, gaming, and other Internet protocols where latency is more important than throughput.  TCP, with congestion control and packet confirmation built in, sacrifices latency for accuracy.  A dropped packet in a phone conversation isn’t much to worry about, but a half-second delay is extremely annoying.  On the other hand, a half-second delay in downloading a computer program isn’t much to worry about, but a dropped packet means that the program won’t run. 

So, as a general rule, TCP runs data apps, while UDP runs real-time apps.  Rudimentary QoS policies based on giving UDP packets higher priority may not be perfect, but they can be a good start to improving performance on simpler networks. 

My concern was that putting BitTorrent, a non-latency sensitive application – on UDP would result in it receiving higher priority traffic.  But after speaking to Simon Morris, the Vice President of Product Management at BitTorrent, Inc., I was assured that this wasn’t the case. Morris explained:

[Editor's Note: Some words got a little garbled in the phone conversation I had with Simon Morris, and he sent me an e-mail with clarifications. I've made corrections via strikethroughs. --ed.]

“BitTorrent obviously needs to be a accuracy-sensitive protocol but we believe it needs to also be …MORE sensitive to latency than TCP, not less sensitive.

This is to say that with uTP, we have taken UDP, implemented a layer of reliability and spent a great deal of time implementing a congestion control mechanism that is better than the one used in TCP (better = faster to detect issues, faster to react).

It’s not QoS, but rather a congestion management mechanism implemented at the end-user’s layer 7 [Application]. This will stop uTP from eating up traffic bandwidth reserved for latency sensitive apps like VoIP and gaming. What’s more, the congestion management mechanism isn’t something we implemented as an afterthought – it’s the whole point of uTP.

In short, it seems that rather than using UDP as a way of getting around TCP’s traffic congestion features, the new protocol is rebuilding better traffic congestion features at layer 7, using the lightweight UDP protocol as a simple base.  In short, to mangle a metaphor, they’re re-inventing a better wheel. 

If uTP does congestion control for BitTorrent as an application, this could provide an answer to BitTorrent critics, and ISPs who claim that BT throttling is necessary because it’s “eating up bandwidth.” I asked how uTP implemented congestion control at Layer 7 [Application] rather than Layer 4 [Transport].

Morris: “So basically, what TCP does is that it stops detects congestion only when it detects packet loss and then it throttles back.  Because we control both ends of the transfer, we can actually measure the single-trip time between when the packet is sent and a packet arrives.  (Not round-trip-times but single-trip-times…)  We have essentially, an ability to monitor single trips over the internet across millions and millions of terminals, and we built an algorithm around that to do things like eliminate the discrepancy between the stock clocksettings on different terminals - to identify where there is actual - very fine grain, down to milliseconds - changes in the speed of which packets are arriving….” 

“The way that prioritization policies are set [by network operators] is extremely varied, and so, it's possible that [uTP-based] BitTorrent traffic will get a higher prioritization, but only in cases where it's not causing any type of congestion at all… [uTP] will never trample over UDP based latency sensitive traffic, nor TCP-based traffic.  [UDP is] designed to throttle back if there's anything else on the line. 

“I mean, it's essentially designed to be - a term that we use internally is - a "scavenger protocol."  It scavanges and uses bandwidth that is not being used by other applications at all, and it's designed to throttle back very very quickly in case there is any type of congestion on the line.

“Unfortunately the way that TCP works is - just profoundly broken as a method of control congestion on the Internet.  Especially when there are applications out there that are designed to get the most out of network bandwidth - like BitTorrent.  People have tried to make TCP better, but the problem is that it's such a huge implementation task to make it happen, because you need to upgrade all of the Web servers and all of the terminals.  Now, the insight here is that in most manycases, we have control of both ends of the [communication].  So we can actually take the right steps in the direction of solving this problem. 

Those concerned about BitTorrent (either classic or newfangled) traffic on their networks might want to check out a solution designed to monitor and track the types of traffic going on the network, including information about what applications are transmitting and receiving what amounts of traffic. 

By and large, the argument over network neutrality tends to be dominated by two specific groups, says Tim Lee. Those coming from a technical background who talk about the important nature of the Internet’s open end-to-end principles, or those from an economic background who talk about how non-neutrality makes business sense.

Lee, a frequent contributor to Ars Technica and Techdirt, has recently written “The Durable Internet,” a paper published by the libertarian-leaning CATO institute. In it, Lee argues that both sides miss a key point – that is, the Internet’s open-ended architecture is not likely to vanish, despite the fears of net neutrality proponents, and despite the wishes of net neutrality opponents. For that reason, perhaps network neutrality legislation isn’t necessary – or even desirable from an open-networks perspective.

In “The Durable Internet,” Lee addresses the concerns in “plain English” but with enough technical detail to argue the point for someone familiar with the makeup of the Internet and familiar with the technical issues involved.

We sat down for a phone call interview with Lee, and you can find the podcast here, with a transcription below the fold.

Interview with Tim Lee, 10.9MB MP3

Continue reading "Network Neutrality without regulation? Interview with Tim Lee." »

Alex Payne, API Lead at Twitter, recently wrote about a blog post beginning “The Internet is built wrong.”

We happen to concur. There aren’t enough pneumatic tubes.

But aside from that, Payne’s criticism is that the Internet, designed years ago for pushing text, research data, and code was designed poorly for the problems we have now – the examples he states are IPv4 – which works well enough but is inadequate for the world’s need for IP addresses, and SMTP, the simple but unsecured nature of which has lead to problems with Spam and DOS attacks on e-mail boxes.

Moreover, he talks about performance:

“You needn’t do more than attempt to watch a streaming video on a busy office LAN or oversubscribed DSL circuit to understand that even the best-served markets for Internet connectivity are struggling to keep up with demand for networked content. Add to this that providing adequate security models for such content is a virtual impossibility on today’s Internet, and the need for a better approach is even clearer.”

The problem is that while Payne makes the case that the Internet doesn’t work as well as “something else” would, and we’re only using the technologies of the Internet because it works just well enough that “something else” can’t replace it even if it’s better – he leaves that “something else” to Van Jacobson and Jacobsen’s idea of “Network Channels.”

Jacobson gave a talk at Google in 2006 on this idea.

The cynics over at Slashdot immediately considered Payne’s idea of a “content centric approach to networking” as some sort of buzzword that means “owned by the media cartels.” However, I don’t think this is what Payne means at all. (If it is, I apologize for not being cynical enough.)

As Jacobson points out, TCP/IP is a very successful technology, but it has a few problems. You can’t connect to things that move – and we’re not talking Wi-Fi. Take trains, for example – as you switch seamlessly from one cell node to another, it’s easy to make a cellphone call on a train. But it’s not so easy to have a continuous Internet session. (In fact, if you do, it’s probably through the cellphone network…)

Additionally, the protocols that were designed for conversations between specific endpoints don’t work as well as they could with broadcasting because the network protocols in use were designed for conversations between two applications on two machines. But Jacobson believes that even the idea of the conversational model isn’t adequate to solve today’s new Internet problems.

From Jacobson’s video:

“We got a chance to look at the data on the routers downstream of NBC’s servers for the Olympics. At one time, their main router got severely congested when Body hit the pole on the slalom. In that router there were 6,000 copies of the same data. Everybody was pulling down the URL. The poor router can’t do anything about – you can’t optimize it, because its dynamic content. It’s all going out in separate conversations. All the router knows is I’ve got 6,000 separate TCP conversations. It’s the same data. If you could broadcast it, you could turn both the router and the downstream links from the server, reduce the bandwidth by three orders of magnitude, but our protocol architecture doesn’t support that. It works at the conversation level…

…Any of the measurements that I’ve seen recently saying that the high 90% level of traffic is people trying to get some named chunk of data. They hand in a URL, and they want to get something back. That’s not a conversation. That’s not a conversational model… that kind of interaction is a dissemination… It’s a point to multipoint or multipoint to multipoint.

Now, this is not to say that the traditional TCP/IP model doesn’t work – Jacobson is keen to point out that the problems we have today with networking only exist because the problems we used to have with networking that TCP/IP solved were solved extremely well. We just have new problems.

Weirdly enough, anecdotal evidence bears Jacobson’s model out. Look at BitTorrent, which is essentially multipoint-to-multipoint dissemination over the TCP stream. And it consists around 50% of the traffic on the Internet. Add eDonkey, another multipoint P2P app, and you get 70%. If that’s not a clear indication of changing demands, I don’t know what is.