By Brian Boyko

For anyone who considers our recent data visualization experiment -- Netcosm --"just a toy," without serious purpose in a large enterprise network, I invite you to take another a look.

I've noticed something incredibly powerful and poignant when asking people what they think of Netcosm, the NetQoS experiment in 3-D network monitoring. By far, the comment most people give me is that "It kinda looks like 'Tron'" - despite the fact that, if you actually go back and look at "Tron" - Netcosm looks nothing like it. The only similarity is a conceptual one - both "Tron" and Netcosm give you the ability to "look inside the network."

But why "Tron" in particular? Why not "The Matrix," which, after all is a much more recent movie. Well, that has much to do with my sampling bias. I had been asking people in my generation - late 20-somethings and 30-somethings. And we all remember Tron from our childhood or young adulthood.

If you ask a young 20-something what Netcosm looks like, they're much more likely to say "It looks like the Matrix." And I've had one or two older Unix gurus tell me that Netcosm "looks like something out of Buck Rogers."

It should not be a surprise that we hold onto the memories of our childhoods first and most dearly.

(Continued...)

Continue reading "Rethinking Network Monitoring: The Re-Juvenilization of IT" »

Information Week: Should IT Get The Blame in Teacher's Pornography Case?

Sharon Gaudin talks about the Julie Amero case and ponders whether the IT department was responsible - and culpable - for the situation which resulted in Amero's trial.

Sophos' O'Brien agrees. "She was the adult in the room at the time responsible for what the children were exposed to, so she became the scapegoat for what was a lack of responsibility on the part of the people who should have been maintaining that PC," he said, noting that the computer in the classroom wasn't running a firewall or antivirus software. "When I look for who is responsible for maintaining that PC, I don't look to the substitute teacher. There was a tool in the classroom that wasn't properly maintained."

And that, added O'Brien, should be a loud warning to CIOs and IT administrators in education, government, and industry. In a corporate setting, pornographic Web sites or pornographic pop-ups on computer screens could lead to office unrest and even create a hostile work environment.

We've written on the network monitoring and IT management implications of the Amero case previously.

Dare Obasanjo: Google Gears - Replacing One Problem with Another.

An examination of the problems with the trend of turning online apps into "offline Web apps" via mechanisms such as Google Gears, Microsoft Silverlight, and Adobe Apollo/AIR, which basically comes down to the classic question: If you make two separate revisions to an old version, how do you merge the changes?

I don't consider myself some sort of expert on data synchronization protocols but it seems to me that there is a lot more to figuring out a data synchronization strategy than whether it should be done based on user action or automatically in the background without user intervention. It seems that there would be all sorts of decisions around consistency models and single vs. multi-master designs that developers would have to make as well. And that's just for a fairly straightforward application like Google Reader. Can you imagine what it would be like to use Google Gears to replicate the functionality of Outlook in the offline mode of Gmail or to make Google Docs & Spreadsheets behave properly when presented with conflicting versions of a document or spreadsheet because the user updated it from the Web and in offline mode?

Don't get me wrong - the "use it even without a net connection" aspect of these projects is great, but I'm concerned that this will inspire app developers to be lazier about how much information they send over the network - it could increase the chattiness of apps. On the other hand, if apps can be used offline, they can also theoretically be used offline when there's too much traffic on the network and update to the online database when traffic lightens. The end-users can keep working during this time without annoying lags in the software. Makes me wish we had more offline apps back when I was entering shopping codes into RETEK for a major east-coast supermarket chain.

WinBeta: Microsoft Study - Vista improves networking.

Taken with the requisite grain of salt, a Microsoft-funded study says that Microsoft's Vista and Longhorn products, used in combination, increase network performance threefold.

In a 38-page white paper titled "Enhanced Network Performance with Microsoft Windows Vista and Windows Server 2008," the Tolly Group compares the networking performance for file access on WANs and LANs using various configurations of Windows Server 2003 R2 or 2008 on the backend and XP SP2 or Vista on the frontend.

Not surprisingly, the combination of the newest operating systems brings the best gains in network performance in large part, Tolly concludes, because of an upgraded TCP/IP stack and updates to thefile-sharing Server Message Block (SMB ) protocol.

We've covered the more-aggressive Compound TCP stack in Vista previously (here and here) and we're not surprised that Longhorn-to-Vista connections would be optimized for one another. But here's the big question: will we need Longhorn to run a Halo 2 server?

We spoke with Jason Potkanski earlier for his work with Wafaa Bilal's "Domestic Tension" project. But since we made contact with him, we found out that he was also the technical director of the expert alternativeto Wikipedia, Citizendium. While we had his attention, we decided to talk with him about the Citizendium project and what it takes to handle such an ambitious Web 2.0 project.

Citizendium is a LAPP (Linux,Apache, PostgreSQL, PHP) configuration, he explained. They went with PostgreSQL for a number of reasons, including better scalability. PostgreSQL is an MVCC database. Unlike Wikipedia, Citizendium never has to lock the database for reads and writes. MySQL can do a lot of things quick and replicate them to slave servers, but PostgreSQL excels at complex functions and full features like JOINs and can do complicated categories and full text searches faster than Wikipedia.

"The reason we went with PostgreSQL was threefold," Potkanski said. "First, to be different from Wikipedia. Second, we already had Greg Sabino Mullane, a core PostgreSQL developer, on board. Finally, we felt from reading various mailing lists over mediawiki development that mediawiki was hitting the ceiling of the features MySQL can provide as a backend."

There is a performance hit, however, with PostgreSQL. PostgreSQL has longer TCP setup times and reduces the amount of users Citizendium can serve compared to MySQL, Potkanski explained.

(Continued...)

Continue reading "A look at Citizendium's Backend" »

Apologies to our regular readers, recent events have delayed publication of the blog. We'll probably have the next post up sometime this Saturday. In the meantime, feel free to use this post to detail some suggestions and ask questions about the blog, about our coverage of the Wafaa Bilal, Julie Amero, Net Neutrality, Vista, Linux in the enterprise - or whatever you'd like. Consider it an open thread.

Oh, and if you've been reading this blog regularly but haven't yet subscribed to our RSS feed, you can now do so easily with one of the chicklets on the left. Let me know if your favorite service isn't covered and we'll try to add it.

-- Brian Boyko

We - that is, NetQoS - recently had our customer service independently evaluated by First Market Research, to try to find out what our customers really thought about us. Apparently, they like us.

In the report, based on a 50-case telephone survey, eight out of ten NetQoS customers had "very favorable" opinions of NetQoS compared to other companies they deal with, citing excellent customer support and high quality products. They also said, overwhelmingly (9 out of 10) that our network performance management products and services have a beneficial impact on their organization, including lower costs, better resource utilization, ability to completely identify the source of problems, greater visibility of network traffic, ease of receiving help, quickly solving problems, and improving application performance.

Additionally, nine out of ten said it was "very likely" that they would recommend NetQoS to friends or colleagues, eight out of ten gave excellent ratings to NetQoS employees for competency, responsiveness, and accessibility. Nine out of ten customers also characterized their investment in NetQoS as "excellent" or "good."

Here's the kicker. 100 percent of customers in the 50-case study said that NetQoS employees are doing at least a "good" job - and 80 percent of them said that we're doing an "excellent" job.

(Continued with the wrestling reference explained, below…)

Continue reading "NetQoS gets high marks from customer service. (Or Why NetQoS would make an excellent professional wrestling company, if we chose to do so.)" »

Just a quick note: Julie Amero, Connecticut Schoolteacher, has had her conviction overturned and will get a new trial.

Here's a link to Brian Krebs at the Washington Post with most of the details.

We've previously covered this, including in-depth interviews with both witnesses for the prosecution and the defense.

• The Strange Case of Ms. Julie Amero: An Introduction
• The Strange Case of Ms. Julie Amero: Commentary by Mr. Herb Horner (Witness for the Defense)
• The Strange Case of Ms. Julie Amero: More Information In The Interlude (Covers Steve DelGiorno, CEO of ComputerCOP software, which featured in the trial)
• The Strange Case of Ms. Julie Amero: Commentary by Detective Mark Lounsbury (Witness for the Prosecution)
• The IT Management Implications of the Strange Case of Julie Amero

Edgeblog: It's still the latency, stupid

Edgeback Consulting talks about how network speed isn't just about bandwidth throughput. It's about latency. Though I don't like being called stupid--I get enough of that from my ex-girlfriend.

I was reminded of Cheshire's article and the underlying principles recently when working on an international WAN design. What Cheshire noted was that light signals pass through fibre optics at roughly 66% of the speed of light, or 200*10^6 m/s. Regardless of the equipment or protocols you use, your data cannot exceed that theoretical limit. This limit equals the delay between when a packet is sent, and when it is received, aka latency.

In the US, we tend to focus on bandwidth and carrier technology when ordering circuits, completely ignoring latency.

If latency was a five year old, it would grow up with severe emotional problems - that's how ignored it is. We agree with this article - in fact, we actually have our own latency calculator as part NetPerformance.com's suite of tools.

Ars Technica: Universities respond to decline in computer science students

Apparently, there are fewer computer science majors out there since the tech bubble burst. Whodathunkit?

It's no secret that there are fewer undergraduate students majoring in computer science today than there were in the late '90s. The Computing Research Association's statistics show that the number of freshman who list computer science as a probable major has fallen by 70 percent since 2000. According to the Associated Press, universities are responding to this trend by attempting to spice up computer science education and make it more appealing to incoming students.

Such as, for example, allowing students to treat college like a World of Warcraft game. For example, there are organizations that are much like guilds on college campuses, called "fraternities" and "sororities." Improvement is done through grueling, repetitive tasks - grinding "homework." And of course, recurring fees. (Many people stick with it a few months but then get bored with it and drop out.)

I have less than fond memories of my own experiences with computer science education. I was frustrated with the emphasis on niche commercial development tools that I had never used before and have rarely used since. I also got frustrated with the emphasis on technical minutiae that aren't particularly relevant to general application development. Assembly programming and compiler design skills acquired in college aren't going to be very useful for software developers who enter the workforce and get paid to write web applications with ASP.NET or Ruby on Rails.

I was a computer science major for two semesters. When I objectively evaluated my chances that the skills I developed learning assembly language specific to a VAX machine would come in useful, I decided to switch majors to history.

ZDNet: Google Gears vies to be defacto tech for Web apps

Online applications are great. The only problem is that they're only as good as your internet connection - the "offline problem" means that if you go offline, you can't access your data or even save your work. Google Gears is a technology (one of many) poised to solve that problem.

The technology consists of three primary components. First, a local "server" which is not really a local Web server but rather a means for capturing all of the Javascript and HTML-based logic that drives the functionality of an application. Second, a local database using the open source-based SQLite for persistence of any data associated with the offline functionality. For word processing, one record could theoretically hold the entire contents of a document. For something like Google Reader (an RSS reader app from Google and, in a proof of concept, the first to take on the offline capability using Google Gears), one record might hold an RSS item. To the SQLite project, Google also contributed the necessary Javascript APIs for accessing SQLite from Javascript.

Thirdly, Google took a look at the single-threaded mentality of most Web browsers and realized that any hangup in a Javascript-based application could easily hang the entire browser (something that has happened to me several times!). So, Google solved that problem by making it possible for Javascript threads to run as background tasks.

My only problem with it: Doesn't "Google Gears" sound like a 1980's toy-based Saturday Morning Cartoon Show?

Last Thursday, we put up a reader poll asking what type of content people would like to see on Network Performance Daily - and while we didn't get any sort of real representative sample, we did get enough people responding to get a very, very rough idea of why people read and what we should focus on in the future.

40% of you - as of Monday afternoon - wanted us to focus on covering Network Performance material - with other information secondary to this main goal. That's a huge chunk, and a majority. However, it would be rude to discount the 30% of people who want us to focus more generally on Enterprise IT, and the 30% of people who want us to focus more generally on technology.

Ultimately, we think this is a good mix, and while we haven't done any content analysis studies to back it up, I'd say that most of the stuff we do for Network Performance Daily roughly falls into that ratio - 40% Network Performance, 70% Enterprise IT (total) and 100% technology.

Of course, as I said, this isn't a representative sample, so we're going to continue to leave the poll up in a sidebar so that readers can continue to let their voices be heard.

Brian Boyko
Editor, Network Performance Daily

*Update on Wafaa Bilal and Recreational Network Traffic

The Brain Behind Paintball Artist Bilal: ‘Domestic Tension’ Sysadmin Jason Potkanski talks about how he helped keep the paintball-via-webcam site running.

Jason Potkanski, who works at the Citizendium Foundation, read the Chicago Tribune Daywatch, to find Wafaa Bilal’s site, Domestic Tension. He looked at the site and saw that it was getting swamped over a DSL link. He sent Ben Chang, lead software programmer of the project, an e-mail offering help, and got an answer back the next day. He visited the project site at the Flatfile Galleries on the 11th – a day when the gun was down most of the day due to the strain that the Internet was placing on the connection – the DSL line and commercial-grade router they used were taking a beating from the Slashdot-like effect of a small pipe.

He immediately offered some performance optimization techniques, such as setting keepalives to 3, and installing eAcellerator to speed up the PHP script, replacing the memory inefficient tail function for the chat. And of course, he told Bilal that he would need a dedicated line to the Internet, and by the next day, they got an OC3-quality connection to Steadfast, at Equinex in Chicago. Immediately he saw an improvement.

He also mounted the webcam image and chatlog/shooter log via NFS, remoted in to do the mounts and copy the application over to the Steadfast connection.

Another problem was the use of XML requests, which wasn’t viable on all browsers, and the fact that the original configuration kept three separate requests open for the chat, the shooter log, and the web cam.

If he had the system to design all over again, he’d design it to use a backend database for the logging features, combine the shooter log and chat so that the application requires only two keepalive connections, not three.

(Continued...)

Continue reading "The Brain Behind Paintball Artist Bilal: ‘Domestic Tension’ Sysadmin Jason Potkanski talks about how he helped keep the paintball-via-webcam site running and Recreational Network Traffic." »