By Brian Boyko

The greatest barrier to creativity is a lack of boundaries. Counter-intuitive - almost zen-like - but we've found it to be true.

And this is why people play Dungeons & Dragons (and similar games), and why network engineers often spend time putting out fires when they could be improving the network.

Allow me to explain.

Dungeons & Dragons, if you're not familiar with it, is a game where people tell a story and when there's a moment of indecision in the game, the players roll dice to determine what happens.

(And yes, I play these types of games, though my favorite is Hero. No, I haven't had a date recently. What's your point?).

(Continued...)

Continue reading "Editorial: Dungeons & Dragons & Networks" »

On Monday, we covered the Daylight Savings Time switch and reiterated the importance of having correct, up-to-date data with which to make your IT decisions. Carol Shiraldi has compiled a short list of engineering disasters that occurred because of bad data.

By Carol Shiraldi

Bad Data: Necessary Height of Bulkheads

The RMS Titanic struck an iceberg on April 14th, 1912 and sank 2 hours 40 minutes later resulting in the loss of 1523 lives. The boat was marketable as "unsinkable" because the ship had been sub divided into 16 watertight compartments by 15 transverse bulkheads. The front six compartments were damaged and flooded, which should not have resulted in the sinking of the ship, except that the height of the bulkheads - three meters, taller than any ship previously built - was not sufficient to prevent water from coming over the sixth bulkhead, causing the ship to sink. After the Titanic sank, the height of bulkheads on other ships was extended even further. The lack of lifeboats contributed substantially to the loss of life, but it was this initial miscalculation that caused the ship to take on water and sink in the first place.

Bad Data: Wind Loads

On December 28th, 1879, the Tay Rail Bridge in Scotland collapsed when a train was crossing the bridge in a gale force wind. The design team for the bridge had massively underestimated wind loads, and the bridge components were over stressed leading to catastrophic failures. The bridge was rebuilt to a similar design with upgraded specifications, accounting for the wind load, and still stands today.

Bad Data: Construction Design

On July 17th, 1981 between 1500 and 2000 people gathered at the Hyatt Hotel in Kansas City to attend a dance. At 7:05, there was a loud crack and the second and fourth floor walkways crashed to the ground. It was the worst structural failure in the history of the United States. The failure was caused because, unknown to the hotel, the contractor did not build what was specified, creating an overload in the attachments and bolts of the structure.

These, of course, are just examples from the fields of engineering. If you look outside of that field, there are many more.

Carol Schiraldi is a Senior Software Engineer at NetQoS.

We invite you to share your own "bad data" stories - in IT or in other areas - through our comments section.

By Brian Boyko

Last Thursday, I made sure to patch my work computer, running Windows XP, for the Daylight Savings Time switch. This past weekend, I made sure to update my TZ database for my Linux home system, as well as patch the computer's Windows partition (for gaming) and the XP session that I run virtually. I also made sure to patch my laptop, and helped talk my friends and family through patching their computer over the phone. I even called my cellphone company to ensure that my cellphone would continue working and display the correct time after the changeover.

But I forgot to set my alarm clock. I thought it would auto-update, but I had confused it with an alarm clock I used when I was a student, and which I no longer have.

I made a decision based on bad data. C'est la vie.

Anyway, the federally mandated changeover to Daylight Savings Time has come and gone, and I hope everything has gone smoothly for those of you working in datacenters and IT departments.

The troubling thing though, is that for all the sound and fury, will the Daylight Savings Time switchover actually accomplish anything, and did the decision to extend Daylight Savings occur because Congress was basing it on a study conducted over 30 years ago - before the multiple-TV and personal computer household?

It looks like Congress made the decision based on bad data, and many people are explaining that there is little benefit to the time change and it may even be counterproductive.

(Continued...)

Continue reading "It's called Daylight Savings Time because they keep you in the dark" »

By Brian Boyko

On Wednesday, Reddit.com, a Web 2.0 site where people discuss the news of the day, was overwhelmed on its front page by calls for the impeachment of President George W. Bush, locking out all other news stories. And the way it happened provides some interesting observations on Web 2.0, 21st century society, and network administration.

A less cliquish version of Digg, Reddit is ultimately very simple. Stories are linked in order of popularity over time. Hit the up button to vote for a story to be placed higher in the rankings, down button to place it lower in the rankings.

From a journalism/new media perspective, sites like Reddit are absolutely fascinating. Before the advent of blogging, journalists used to talk about gatekeepers – reporters and editors who choose what to cover, how to cover it, and what emphasis it is given. A story on the front page is given more emphasis than one buried in the back pages – and therefore, more people will read it and more people will think it is significant.

Blogs and blogging took this to the next step in that anyone could become a gatekeeper by publishing their own blog and figuring out what they thought was important. Readers also could act as their own gatekeeper and simply go to the stories that they think are important. It removed the gatekeeper as the arbiter of importance.

Sites like Reddit are a logical next step. In blogging, the blogger takes on the gatekeeper role – but it is still as much an elitist position as far as the blogger writes about what he or she thinks is important. Reddit and its Web 2.0 kin allow for the gatekeeper role to be supplanted by a reasonable facsimile of direct democracy.

And what you have is a situation where the wisdom of crowds takes hold. Who better to determine what “the people” find important than the people themselves?

(Continued...)

Continue reading "Reddit’s Tea Party" »

By Brian Boyko

Recently, we had a chance to speak to Joe Miller, VP of Platform and Technology Development at Linden Labs, the makers of online “virtual world” Second Life, regarding their plans to rollout VoIP on the popular world, known for being the online virtual homes of businesses (such as Cisco), news organizations (such as Reuters), and political campaigns (such as John Edwards).

According to Miller, the VoIP product is unique because of the ability to project the sound in three dimensional space, as a function of distance and direction from one avatar to another. It takes a 32khz signal at 32kbps from clients, sends it to an Intel based audio server where the input signals are mixed and properly positioned, acoustically, in three dimensions, and a stereo stream is sent back to the client at 64kbps. Even with 100 people speaking at once, the bandwidth requirements are the same for each individual because the servers (dual quad-core Xeons) mix the voices together into a single data stream.

The codec used is Siren 14/G.722.1 Annex C, developed by Polycom but now an international standard. It was chosen because it uses relatively low bandwidth but can carry a wide and dynamic range of audio – not just human voices – making it an ideal codec to broadcast, say, a musical event.

“It carries the complete human hearing range – it’s not just a speech codec, it carries music and ambient sounds quite well.”

(Continued...)

Continue reading "Joe Miller of Linden Labs speaks about VoIP in Second Life." »

By Brian Boyko, Editor, Network Performance Daily

Network Performance Daily is my bread and butter, but I sometimes freelance at HardOCP.com (or "[H]ard|OCP"). I used to work there full time as an OEM computer evaluator - we'd be "reviewers" only it's hard to call a 10,000 word article a "review."

Today, one of the articles I wrote for HardOCP was published. This article, "30 days of Ubuntu Linux" was an in-depth, massive article complete with pictures and, at one point, video, of Ubuntu Linux from the perspective of the end-user. And, well, it's a "hit."

I'm very proud of it, obviously. The story was published at around 9:00 a.m. this morning, landed on Slashdot at 9:29 a.m. and on the front page of Digg at around 11:00 a.m. I'm still trying to deal with the fallout of it all.

Now, HardOCP has been around quite a while - it's one of the major computer review Web sites. HardOCP has tons of bandwidth - they are used to very high traffic every day.

That said… the site went down. [H]ard.

We had our own little version of March Madness. I'm perversely proud of that.

Of course, user Mink78 at Digg.com pointed out that my "network performance needs some work." No - the irony is not lost on me, and I'm sure it's not lost on HardOCP, which, after all, sells advertising space. If the page can't be loaded, neither can the ads.

So, while HardOCP's problems with serving pages are problems that they're happy to have, I'm sure they'd be even happier if they didn't have them. Web 2.0 sites have the potential to make a story amazingly popular very, very quickly. Traffic often comes in spikes which are rarely predictable - so you need to be able to know what your network can handle, for how long, and compare it to normal.

By Brian Boyko

In some ways - not many, mind you, but some ways - the days of 14.4kbaud modems were a golden age.

14.4kbaud isn't much. By today's standards it's not even acceptable for dialup speeds, but it was enough to allow connectivity and communication worldwide. It allowed us not only to participate in worldwide conversations, but these conversations were happening, for the first time, not just among the technical elite and the academics but regular people.

This was the era of Usenet, Listservs and E-mail, of IRC (and AOL chat) and the very beginnings of the popularized World Wide Web. And all these communication mediums had one very simple thing in common.

They used text.

ASCII text, to be precise. Because text is a great medium for communicating when you have very little bandwidth.

Because of that, the popular Internet evolved, primarily, around text, and different ways to get that text from one user to another. And so, for most of the previous decade and of this one as well, text was the dominant communication medium on the Internet.

(Continued...)

Continue reading "Friday Editorial: The Internet Paradigm Shift" »

By Brian Boyko

Although the stock market seems to have started recovering, trading in the New York Stock Exchange took a heavy tumble yesterday.

There are two main theories about the market plunge. The first was a $100B drop in the value of the Shanghai stock market.

The other is computer related. According to the Seattle Times, Dow Jones & Co.'s computers weren't up to the task of handling the huge volumes of trades at the NYSE - a spike of 4.5 billion trades, or twice the usual average. Switching over to a backup computer caused a massive "catch-up" in the Dow Jones industrial average - which recorded an instant plunge of 200 points, which couldn't have had a positive effect on investors, who reported that computer systems were slow to respond.

(Continued...)

Continue reading "Performance Problems Lead to Drop in the Dow" »

By Tim Smith, John Mao, and Ben Erwin. Compiled by Brian Boyko.

Dale Imler wrote in:

I would like to see some information on how to use the SuperAgent and Reporter Analyzer for executive reporting and capacity planning.

The big consideration with capacity planning is that it shouldn't be done in absence of performance considerations, and you can't consider performance with insufficient data. Mostly, SuperAgent helps to positively identify those cases where the solution to intermittent performance is increased capacity, and help eliminate false positives where the solution might be something completely different.

Additionally, SuperAgent can measure the degree in performance gain you can expect following an infrastructure upgrade. If you have WAN circuits of various types, SuperAgent shows application performance over each type, and may reveal patterns associated with that type.

If you're using some sort of WAN optimization, you should keep performance in mind when making utilization based capacity planning, because some types of WAN optimization are designed to raise utilization rates in order to extract better performance. So, if WAN optimization is present and performance is excellent, high utilization needs to be interpreted differently than if WAN optimization wasn't there. Again, this means measuring end-to-end performance with a product like SuperAgent.

(Continued...)

Continue reading "In response to Dale Imler's question..." »

by Dr. Cathy Fulton

WAN optimization devices are certainly great pieces of technology. They do much to alleviate the problems that companies have with keeping applications which travel over the WAN responsive. But deployment is not without risks and companies need to look at cost vs. performance benefits.

There are several benefits to WAN optimization. One is offloading WAN links – when there is less congestion every application performs better. Another benefit is minimizing the impact of latency on the network. Even if it doesn’t reduce the traffic on the network, TCP and application optimizations can reduce the effective latency experienced by your data.

You might even get a benefit from increasing WAN utilization while still receiving good performance - by fulfilling a latent demand. A good example of this would be a company that receives orders over a network. With WAN optimization, the holy-grail scenario is that people who had been frustrated because of poor performance would now place orders on a more responsive order system. You have to have exactly the right circumstances and environment to hit the points where utilization goes up while end-to-end response time decreases, but it is certainly possible.

But WAN optimization is not a panacea. It works well on a subset of performance issues, but can actually be detrimental for others. In terms of mitigating the effects of congestion, it works best on small to mid timescales; it certainly does not eliminate the need for longer timescale techniques including capacity planning. And it does not work well with every application – some perform worse when “optimized”. If you’re interested in what sort of benefit you’re going to get from your WAN optimization efforts, you need to look at the amount of congestion on your link and the composition of that traffic.

(Continued...)

Continue reading "WAN Optimization Devices: What you need to consider before deployment." »