When Cisco Live! (a.k.a. Cisco Networkers) starts up next week, every vendor is going to try to get you to their booth, learning about their products, and telling their stories. We’re no different.

I would like to take this time to point out that all the cool kids are going to booth 807 this year, and the more time you spend at booth 807, the cooler you are.

Now, anybody can make that claim about their booth. For example, we just did.

In all seriousness, I am really looking forward to the presentation on Tuesday, June 30^th at 2:30 that NBC Universal is going to be giving along with Steve Harriman, our Senior VP of Marketing, at Cisco Live, because while there are “swagger stories” abounding in this industry, sometimes some stories are just too cool.

In this case: How do you send live coverage of the Beijing Summer Olympics through to television viewers, cable viewers, mobile device viewers, and Web viewers – data totaling 3.4 petabytes?

3.4 Petabytes is over three and a half million gigabytes. Put into context, if you put that on 1.44MB floppies, the height of the stack would be roughly 2500 miles - equal to the distance between New York and Los Angeles by air. In order to pull it off, NBC had to use three 155Mbps OC-3 pipes to get the footage from Beijing, of course, monitoring and maintaining the network the whole time.

There are plenty of stories about optimizing your network, but that’s a story I really want to hear – it’s a big task, no room for error, and only one shot to get it right. Plus, you know, it’s Television. Say what you will about mocking the “old” media in the age of the Internet, television has always been glamorous, exiting, and even today remains the best way to reach the most people. So, I’m pretty psyched about it.

It’s not a particularly controversial statement that most of what we go through at the airport in the name of “security” is in actuality, “security theater,” as security expert Bruce Schneier wrote in “Beyond Fear.” Security theater measures are measures that don’t actually have any real effect or purpose towards keeping us safe, but they seem like they might, so we go through them anyway, and pay for them.

Whether or not the theatre creates a discouraging effect on attack attempts is debatable, and how much restriction on people’s behavior without any actual tangible benefit is acceptable to maintain the psychological discouragement, if any, is also very controversial.

What isn’t debatable is that airport travel is annoying as hell. The No Fly List has expanded tremendously, even for those of us lucky enough not to be on the list, we have to take our laptops out of our backpacks or briefcases and scan them separately, even taking off our shoes. Some airports are using a backscatter X-ray which means you’re essentially posing nude for the camera every time you walk through.

To quote Janice from The Muppets, “Listen, I don't take my clothes off for anyone, even if it is ‘artistic.’”

Anytime I quote the Muppets, it’s a sure sign I’m starting to get off-track. Anyway, where there’s annoyance, someone will try to make money by selling convenience, and airport security is no exception. In this case, it was the “Clear” lanes at airports, where you would pay $99 per year for a card that signified that you were a minimum security risk, and therefore you could be processed through airport security more quickly.

(Apparently, the main qualification for being a “minimum security risk” is being rich enough to blow $99 on a membership that gets you through lines faster. It also helps if your name isn’t Al Kyder or Terry Wrist.)

Despite some snags, such as “Clear” passengers unable to keep on their shoes or jackets, as they had hoped when Clear launched, Clear continued and it is the business success that it is today because of a savvy—

--Hold on, someone just handed me a note.

Ah. They failed, and are shutting down today.

Which brings me to the idea of doing QoS incorrectly. Yes, QoS priorities can help with making sure mission critical data gets there first, but stuff that isn’t mission critical, merely “mission nice-to-have” can get bogged down as well. If there is some external reason why every packet on the link seems to be performing slowly, maybe it’s a good idea to figure out what that reason is.

One of the reasons for long lines at security in airports (which themselves are a security risk – a suicide bomber could take out a planeload of people without even purchasing a ticket if he blew himself up in the security queue) is because of this security theatre. If that can’t be shown to be effective, maybe it’s time to do something different.

On the network, it doesn’t make a whole lot of sense to prioritize packets until you look at the entire network, from end-to-end, and figure out why the network is slow. Sometimes QoS prioritization is the answer. Other times, it may just be an expensive mask to a more deeply rooted problem.

Earthweb’s Bruce Byfield recently wrote about the innovations in KDE 4 and GNOME; the two main Linux desktop platforms.  More specifically, he wrote that the upcoming versions of KDE and GNOME were designed with numerous improvements that developers decided to put in because they were “cool” – but which may overcomplicate things for the end user on the desktop, using Linux as an alternative to Windows or Mac as a workhorse machine.

(For the uninitiated, KDE and GNOME are desktop environments that provide a unified “look and feel” among graphical applications running on Linux.  It’s similar to the relationship between the Windows XP “look and feel” and the Windows Vista “look and feel” – both do the same thing, but look different and have different functionality.  Each desktop environment also provides tools for developing standalone applications – while applications designed to run in KDE will run in GNOME, and vice versa, they are likely to look and feel slightly out of place.  The difference is, while XP and Vista are developed by the same company, sequentially, KDE and GNOME are developed by different teams of developers, in parallel, and end-users choose the desktop system that meets their needs. )

The theory is that open-source developers who aren’t working for a paycheck are instead working to either add the features that they think are personally cool, and what they think they’d like to see in a desktop platform.  But end-users are more likely to prefer simplicity and familiarity.  While “doing new things” is always a bonus, more people would rather focus on using the computer to do the things they already know how to do, and any “new thing” which gets in the way of that is likely to be seen as an annoyance. 

From Byfield’s article:

“What innovators and early adopters can easily forget is that they are a minority. Where they are excited by change, most users are uncomfortable with change. Many will reject any change out of hand, no matter how logical or convenient, simply because it is new….

…But what the innovators are forgetting is that, for the average user, the desktop is not the destination. Nor is the destination even the application…. Rather, the destination is the user's purpose: finishing the quarterly report or IMing a girlfriend. As they focus on the task at hand, users may not want to linger on the desktop to play with its features.”

This is one of the reasons that we often harp on having network tools that can provide the necessary information to the right audience.  The network engineer is concerned with throughput and latency, the CIO more concerned with how much money good throughput and low latency save, if any. 

Real human usability factors are often as important, or more important, in developing software products – any software development company often has specialists in user interface designing the look and feel of products – NetQoS is no exception.  But user interface design isn’t just about making the functionality “look pretty” – it’s finding better ways to present information – even if that means finding a “less clever” way to present information in favor of a “more parseable” way to do so. 

I don’t know if there’s as much of a disconnect between IT and the business as there used to be; I think that over the past few years, more executives see the value of IT because more people in IT are explaining the value of IT in terms that matter to business executives.  But the quickest way to encourage that disconnect is by giving the C-level executive information he neither needs nor cares about – it’s a quick way for him to tune out anything important that you may have to say.  

Cisco Live! a.k.a. “Cisco Networkers” is going to start the week after next on June 27^th, and I thought it might be time to start plugging our appearance at the—wait a minute…

“Cisco Systems.”

Based in San Francisco.

Oh my god.  I just got that.

I feel like such an idiot now.  Really – I thought it was named after someone like a John Cisco or something.  Like he was the founder of the company.  I’m probably thinking of John Chambers, the current CEO, but the company was founded by Len Bosack and Sandy Lerner.

What else am I missing out on?

Anyway, NetQoS, as “diamond sponsor” of the event, we’ll will have a “Super Session” on Tuesday, June 30, at booth 807 at the Moscone Convention Center in San Francisco, and--

I’m not going to be able to get that out of my head now.  It’s like seeing the arrow in the Fedex logo.  What else have I been missing in my life that’s so obvious to the people around me?  I feel like Grover, who found out that the monster at the end of the book was him all along.  (Him and his drinking problem.)

I’m going to need some time to think about this. 

Well, coming from the land of the ice and snow, from the midnight sun where the hot springs blow, the Norwegian based Opera Software is following in the footsteps of the Vikings.  Opera Unite is a technical achievement and if it – or a worthy imitator - is widely adopted, will be a game changer.  The Vikings were the game changers of their day, engineering the best seafaring technology of the time.  They also pillaged and burned a whole mess of Europe, which many people may find reprehensible, but, hey, the Vikings were the best at what they did. 

At its core, Opera Unite is nothing more than a webserver – we’ve had these things for years, of course.  What makes it different is the ease at which it can be set up – download the browser, create an Opera.com account, and you’re done.  By routing through Opera’s servers, you don’t need to mess with VPN, Remote Desktop, IP addresses, or configuring those fiddly little port forwarding settings on your home router – you don’t even need to have access to it – or even to know what a router is – to use Opera Unite.

Right now, the immediate use is to run it on the home computer to access files from work, and vice-versa.  And that’s where the snag comes in.  Opera Unite cannot be considered in any way “secure” – the fact that it connects to a third-party server makes it ripe for a man-in-the-middle attack, people might mistakenly share sensitive information on their work or home computers, and of course, there’s the problems that you’d expect to have with any filesharing app on the PC.

But more importantly, this will have a major impact on performance.  Employees running data servers can choke the network links your company pays for; we have already seen this when highly technical users run FTP or Web servers from their office desktop machines.  Opera Unite doesn’t change the nature of FTP servers – it just places creation and access of FTP servers and hosting services in the hands of the many, rather than the few. 

Even so, I can’t help but think that overall, Opera Unite is a good thing overall.  It reduces dependence on third-party hosting sites and cloud apps – like YouTube.  If there’s a video that YouTube doesn’t like – or has to take down because of a mistaken or fraudulent copyright claim, the video can still be made available.  (Often times, copyright law is abused to get hosting services to remove unflattering footage of a company or organization.)  And if a particular hosting service should die, the data can still be accessed.  In this manner, it removes some of the risk from cloud computing by allowing anyone to run their own hosting services. 

It doesn’t hurt Opera that they’ve found a hail-mary pass for desktop software to remain relevant in the age of the “software as utility” philosophy of cloud computing… by essentially providing a desktop app that turns your desktop into a cloud computing platform. 

Additionally, although I don’t hold out much hope for it, it may increase demand for more uploading capacity for home users.  AT&T DSL, for example, maxes out at 768kbps upload speed – over three hours to upload a single gigabyte of information. There’s been little complaint about this because most people care more about download speeds than upload speeds – with their own hosting services however, people might be more likely to notice, and care, about what speed they can access their home computer from work, or their friends or colleagues can access their files.  That may lead to increased upload capacity provisioning. 

Ultimately, I’m psyched about Opera Unite as a desktop user, but in the trenches of IT, I’d want to make sure I had a way to track this traffic and see if it affects network performance in a meaningful way.  It may be nothing but a flash in the pan, but if it does catch on, I’d rather err on the side of caution.

Recently, Chandra Hosek, who maintains the NetQoSLive Twitter account located to the left of this blog-post, unless you’re reading it sometime in the far future after which we may have redesigned the site meaning it could be anywhere on the page, or even replaced by a service even more twitter-like than Twitter, or perhaps you’re reading this from a space station and you’re a robot.  A robot that can feel love, and question its existentia—

Where was I going with this?

Oh, right.  (That’ll teach me to get 6 hours of sleep over 48 hours.)

Well, recently, Chandra started asking people what some of their favorite network engineering blogs were, so that we could not only add them to our blogroll on the site, but maybe there’s a blog out there that’s really got some relevant information and we’ve missed it. 

I know – obviously NetworkPerformanceDaily.com is your “favorite.”  But we’re looking for vendor neutral blogs, specifically, so NPD wouldn’t qualify.

Right now we’ve got a couple – there’s The Network Zone, which has a great post up on how to put value in your network where it matters most, and Should Have Gone With Cisco, which recently put out a walkthrough on how to configure the Cisco Nexus 7000 router virtual portchannel feature – but we could use more.  Could you shoot us some links in the comments section of this post?  Thanks!  

There have been many jokes about the irrelevance of Twitter – or at least about the irrelevance of much of the content on Twitter. Some of them have been amazingly creative and funny, like the Tonight Show’s “Twitter Tracker,” and some of them not funny at all, like any of the jokes we’ve made on this blog on the same subject.

But the events of the past few days have killed Twitter irony. In a country where “subversive” blogging is punishable by death, Twitter (along with YouTube) has been the go-to to get information about the happenings in the tightly-controlled Iran.

Furthermore, the relevance of the newest of the new media highlights some of the irrelevance of the “traditional” media. The Revolution, it seems, will not be televised – CNN.com didn’t mention the unrest for days. As ReadWriteWeb.com put it:

“Hours after Iranian police began clashing with tens of thousands of people in the street, the top story on CNN.com remains peoples' confusion about the switch from analog TV signals.”

CNN, the TV Station, provided only regular news reports instead of wall-to-wall coverage akin to the coverage of the Tiananmen Square protests or the first Gulf War – the events which practically made the CNN news channel. If 24 hour news won’t provide 24 continuous hours of news coverage on the most important subjects, what, pray tell, is the point of 24 hour news stations?

Instead of covering Tehran, CNN showed a rerun of the Larry King show, where King interviewed the stars of “American Chopper.” Now, I get it, sometimes newsdays are slow, and sometimes you need to fill in the gaps. But this was anything but a slow news day.

Now, you could make the argument that Twitter is reporting, among other things, rumors and mistakes. If the information coming out of Iran is accurate, CNN just got shown up by a better news service; if it’s inaccurate, CNN should have been dispelling those rumors.

In contrast, Twitter (the company and service), recognizing the importance of the news from Iran, delayed it’s scheduled downtime so that it can remain available for those Iranians who can still access it.

In the meantime, Iran’s government is doing everything they can to prevent the news from getting out – a futile effort in most cases. Looking at the Iranian Internet services, you can a clear pattern of additional outages and unstable connections – starting on Saturday. (Kudos to Renesys to making this information available.)

Of course this is encouraging for encouraging the promise of Democracy. For good or ill – in this case, good – it’s extremely hard to fully block comment and communication on the Internet. There are still sysadmins out there who think that blocking YouTube is an effective response to over-subscribed enterprise networks. Iran, a dictatorship, with an army and a nuclear program for crying out loud, can’t block YouTube completely – what makes you think you can? This effectively illustrates why the appropriate response is shunting YouTube traffic into a lower quality of service so that it doesn’t interfere with business applications, and running network traffic analysis to make sure that it doesn’t.

Iran’s governmental system is interesting because unlike many other world dictatorships, it sets up an expectation of democracy; the idea being that concepts of voting, parliament, democratic representation – they’re not only not foreign to the Iranian culture, but, as we can clearly see from the protests, Iran has one of the most – if not the most – vibrant democratic cultures in the Middle East, in complete contrast to having one of the least democratic governments in the world. Which makes the YouTube and Twitter coverage extremely important - what many Americans are learning from it is that Iran is not a country of extremists and radicals, but a modern, progressive nation with a repressive, barbaric government.

I wish ‘em luck.

When Savvis promises “proximity hosting,” they mean it – according to this New York Times Magazine article. In Weehawken, New Jersey, right outside of the Lincoln Tunnel, there’s a data center that houses the Philadelphia Stock Exchange’s computers. (The PSE is now part of Nasdaq.) Firms compete to have their computers located close – physically and in the networking sense – to the trading exchanges in that data center. Milliseconds of latency are unacceptable in this environment.

“It used to be that things were done in seconds, then milliseconds,” Varghese Thomas, Savvis’s vice president of financial markets, told me. Intervening steps — going through a consolidated ticker vendor like Thomson Reuters— added 150 to 500 milliseconds to the time it takes for information to be exchanged. “These firms said, ‘I can eliminate that latency much further by connecting to the exchanges directly,’ ” Thomas explained. Firms initially linked from their own centers, but that added precious fractions of milliseconds. So they moved into the data center itself. “If you’re in the facility, you’re eliminating that wire.” The specter of infinitesimal delay is why, when the Philadelphia Stock Exchange, the nation’s oldest, upgraded its trading platform in 2006, it decided to locate the bulk of its trading engines 80 miles — and three milliseconds — from Philadelphia, and into NJ2 [in Weehawken, NJ], where, as Thomas notes, the time to communicate between servers is down to a millionth of a second. (Latency concerns are not limited to Wall Street; it is estimated that a 100-millisecond delay reduces Amazon’s sales by 1 percent.)

Back in March 2008, electronic trading made up 60-70 percent of the daily volume of the NYSE. (I’m sorry I don’t have more recent numbers, but they might have been artificially affected by the credit crisis anyway.) And when you remove human beings from trades; the only thing that matters is the speed of a sale; whichever seller’s computer connects first to the buyer makes the sale, whichever buyer connects to the low-bidding seller first gets the bargain. Speed, while not everything, is not underestimated – and it’s one of the reasons you need to identify immediately any problems with network performance in financial applications. Every second a problem doesn’t get fixed – even problems that are imperceptable to the end user, like an added 3ms of delay - means more money is lost.

Now, if your company is over-leveraged and built on shaky investments, network performance won’t save you – we’ve seen a lot of companies with very good network infrastructures go downhill these past few months.

If you want to learn more about the topic of monitoring trading applications for performance, you might want to check out Alex Malone, Software Engineer Manager at NetQoS, who will be speaking at the Securities Industry and Financial Markets Association Technology Management Conference & Exhibit on June 23-25 in NYC. Alex is scheduled to speak June 24, at 2:35pm. You can also look us up at booth #1822.

I got some good news today. Hulu.com is putting the first season of my favorite TV show, ever, “The Mole,” online. (Seriously. Watch it. It’s good.) There is no doubt that video on the Internet has left the realm of novelty and has entered the mainstream; many people watch as much or more “tv” delivered via the Internet then people actually watch a physical television. Some, like myself, only watch video downloaded via the Internet.

The upshot of all of this is that Cisco recently updated their Visual Networking Index which tries to forecast where the Internet will be in the future. And it’s breathtaking.

• In 2008, the Internet transferred 9 exabytes (1 exabyte = 1.074 billion gigabytes) worth of data each month. Cisco predicts it will carry 56 exabytes per month by 2013. This means that annual traffic will reach around 2/3rds of a zettabyte (1 zettabyte = 1024 exabytes).


• The boogieman of years past, P2P traffic, will continue to grow in absolute terms but decline as a percentage of Internet traffic. Internet video, now 1/3^rd of all consumer Internet traffic not including P2P file sharing, will account for over 60% of all consumer Internet traffic by 2013. Including P2P, TV, and Video on Demand, this will account for over 91% of all consumer traffic by 2013. Video communications traffic growth is small but accelerating, real-time video is growing in important, and VoD traffic will double every two years through 2013.


• Mobile phone data traffic will also double every year through 2013, and 64% of that will be video data – which gives a lie to the idea of calling it a mobile ‘phone.’


• While today, business Internet traffic is twice the volume of WAN traffic, but video on the WAN will result in business WAN traffic growing at a faster pace than business Internet by 2013.


• Twitter, Youtube, and Facebook will merge to a single social networking site, called “YouTwitFace!” – No, wait, that one was from the Tonight Show with Conan O’Brien.

One of the traditional solutions to solving online video problems in Enterprise environments has been to block all Internet video entirely; as businesses will require more video information to collaborate between other businesses and with customers; this becomes a less viable option. The trick is managing network video to make sure it co-exists with your data applications – and the sooner this happens, the easier it will be in the long run.

Earlier, we wrote about how March Madness affects Internet performance every… well, March, and as such, it was such a predicable event that IT pros should be aware of the effects of the tournament by now. This growth in video is no different, it’s just constant – a March Madness every day.

This means that if you haven’t already prepared for network service in the era of Net Video, it’s time to step up, think about whether or not to filter packets, what to filter, how to protect your data applications, know which applications need which quality of service, and understand the differences between quality of service tiers. Five years from now, handling Internet video will be the main priority of the network engineer’s job; so it’s a good idea to study up today.

A quick summary of theAberdeen Group Benchmark report, “Getting IT on the C-Level’s Agenda,”:

It ain’t.

Well, it’s not on the agenda of most companies and Ann Bednarz at Network World explains how bad the situation is – despite years of hammering home the importance of identifying problems before users are effected, 61% of the respondants in the Aberdeen study said that they couldn’t do that. Despite the importance of measuring the business impact of application performance, 49% said they couldn’t do that. A full 46% even said that they lacked visibility into the end-user experience.

And the companies affected by these problems cited decreases in employee productivity and customer satisfaction. Conversely, those companies that took steps to address these problems reported an 83% success rate in preventing issues with application performance before users are impacted, with 113% average improvements in application response times.

That companies have problems in these areas isn’t unusual; but the percentage of companies that haven’t addressed these problems this late in the IT game astounds me. For example, take this quote from the Network World article:

“Originally we thought that we didn’t have enough bandwidth to run an ERP application and were thinking about adding an additional T-1 line. However, a technology solution that we had in place helped us realize the root-cause of the problem was on the application side -- not the network,” the IT director shared with Aberdeen. “This allowed us to avoid adding more bandwidth, but it also allowed us to be able to effectively manage application performance on an ongoing basis and address potential problems before they could impact end-users.”

If there’s one thing that we’ve been trying to hammer home with this blog, (other than that D&D is awesome and that you shouldn’t knock it until you try it,) it’s the idea that without visibility you’re not going to be sure where network problems lie, and you’ll end up wasting money on guesswork solutions which may not actually solve the problem in the first place.

More importantly; having visibility into your network and application performance allows you to make a very shrewd business decision. Getting “the best possible performance” out of your network is good, but in reality, most often enterprise IT is tasked with finding a baseline acceptable performance, and finding the lowest possible price that it takes to achieve that performance. Of course, you can’t tell what acceptable performance is unless you know what “normal” baseline performance is, and you won’t be able to tell what’s acceptable unless you have visibility into the end-user experience.