By Brian Boyko

With all the (deserved?) hype over virtualized servers, and the consolidation of hardware in order to reduce underutilized servers or replace obsolete hardware, many are overlooking what might be, in retrospect, an obvious use of virtualization – the end-user client computer.

The benefits of virtualized desktops mostly deal with administration and reducing problems caused by user error.

First, virtual configurations are essentially files on the hard disk, and those files can be backed up and restored to the hard drive of the end-user. It’s not a secure form of data storage, but it can be a lifesaver when a malware infection strikes.

Second, it can shorten the “test and deploy” period for software updates, because virtual environments can be configured to be identical and even in a worst case scenario, where a deployed change manifests problems, you can restore from the aforementioned backup.

Third, multiple VMs can be loaded on one machine. This means that if a particular program hangs or crashes, or works fine but is a notorious RAM hog, the end-user can switch over and continue working on a secondary virtual environment until the IT team can get to solving the original problem.

A possible unintended side-effect, however, is decreased network performance – virtual machines simply don’t run as fast as full-fledged installs, and if you have the ability to back up every end-user’s configuration to the network, eventually you may find yourself having to do so. Those with consolidated data centers and a lot of traffic operating over a WAN might not find the prospect of backing up hundreds of computers, including operating system, drivers, applications, and data, so appealing.

Kept on the LAN, however, servers dedicated to doing backup should not be as daunting. Those servers can be maintained via SSH from the WAN, and would not need that much hardware. You could probably repurpose some of those underutilized servers you have lying around – the ones that became redundant when you rolled out virtualized servers.

WebPerformance: Load Testing a Virtual Web Application

Christopher Merrill takes an academic look and tests out the performance of virtualized servers. He notes that for the added ease of server administration, you are taking a performance hit, and can serve a fewer number of users.

We ran 4 load tests on the application. The first measured the performance of the web application running on a native Windows 2003 Server installation. The second test measured the performance of the web application running on a virtual Windows 2003 Server installation running within VMware Server 1.0.1. VMware ran on Linux (CentOS 4.4). The base hardware for both machines is a Dell Poweredge SC1420 with dual Xeon 2.8GHz processors. The virtualized server has the same memory available to it (2G) as the native server (which implies that the physical machine running VMware has more memory). Since the Intel Xeon processor supports hyperthreading, there was some debate as to the impact of hyperthreading on the virtualized machine. So we decided to run two additional tests with hyperthreading disabled for both the native and virtualized servers.

Merrill makes a point of stating that his own company transferred 7 underutilized servers to two machines with newer hardware and found an increase in performance and productivity – but that was mostly because those servers were underutilized. This is a test of peak performance.

Network World: Layer 8 – Cisco develops communication relay robots

Never a wireless access point around when you need one? Cisco may have solved that problem, albeit in a spooky way.

Cisco has developed a set of small smart robots, which can act as wireless communications relays, that sense when a mobile user is moving out of service range, and can follow the user to maintain connectivity…

Whether or not the systems has an enterprise application, it is of interest to the military and initiatives such as the Army's Future Combat Systems which uses a variety of advanced systems to achieve battleground superiority.

One thing is for certain. There is no stopping them. The Cisco robots will soon be here. And I, for one, welcome our new wireless robot overlords. I’d like to remind them that as a trusted blogger, I can be helpful in rounding up others to toil in their data mines…

When most people think of running Windows programs without Windows, they usually think of WINE, the API developed for Unix systems. But there's another project, called ReactOS, which is an open-source, GPL-compatible operating system designed to run Windows programs and work with Windows drivers - essentially, rebuilding Windows from the ground up without using Microsoft's code. ReactOS and WINE share quite a bit of code and contribute to each other's products, but Unix has its own TCP/IP stack. ReactOS's TCP/IP stack had to be built for ReactOS.

Art Yerkes's day job is working with Secret Level to work on the remake of the classic Sega Genesis hit "Golden Axe." But he's also known as the developer for much of the network code of ReactOS. We had a chance to talk with him, briefly, about his experiences.

"In a previous life, I had done a lot of work on media streaming, and so I got to be pretty familiar with TCP/IP itself during that time. I figured that much of what we had was good up to the packet layer, so I just started pulling the string on stuff I didn't really like. It started at the high level TCP and AFD implementations which were really kinda hokey and didn't respect IRQL at all. There are certain rules to coding NT drivers, and while I wasn't familiar with them at first, the [ReactOS] kernel [developers] people really helped me get up to speed and helped me write the drivers."

Yerkes's work didn't come in a vacuum. He explained that there was a partial network stack already in place, coded by Casper Hornstrup, but that it needed improvement.

"I imported TCP from the FreeBSD snapshot from OSKit… and I kept for the most part Casper's implementation of the packet-oriented protocols, but I cleaned them up and made the implementations a lot smaller and a bit sturdier."

While Yerkes improved the stability of the TCP/IP stack, he admitted that at this point, he's not concerned with transmission efficiency.

"It's much too early in ReactOS's development to think about that. The [operating] system, as a whole, isn't very stable. Leaving the TCP code in as understandable a form as possible is very desirable, if for no other reasons than getting a good stack trace when something breaks, so you can tell where the failure came from."

Yerkes urged newer developers interested in learning how Windows networking works to join the #reactos IRC channel on the freenode network to talk with people there about the project.

Correction: Art Yerkes name was incorrectly spelled "Yerkis" twice in this article. NPD regrets the typographical errors.

Burton Group: Don’t Wait for Windows Vista to Optimize WAN Performance; Do It Now

The Burton Group talks about Vista’s WAN performance improvements, but recommends “enterprise IT organizations currently evaluating WAN performance solutions purchase products now rather than wait a few years for the widespread installation of Microsoft Windows Vista.” Their chief reasons are because new operating systems can take a long time to fully deploy and WAN optimization devices can pay for themselves – according to the Burton Group – in only a couple of years.

Siegel reports that when wide-scale adoption of Vista arrives, the major performance changes will be in these three areas:

• TCP/IP stack: notable performance increases for high-bandwidth or noisy communications paths and for multiple-CPU or multi-core systems because of flow control, error recovery, and scalability improvements
• File services: faster access to remote files because of improvements to the Common Internet File System (CIFS) protocol
• Quality of Service (QoS): enhanced ability of administrators to manage QoS because of improved control over QoS markings

There are a couple of optimization techniques that can’t be implemented at the operating system level – caching, for one. Still, anything that improves network performance isn’t a bad thing – assuming the rest of the OS is up to snuff.

I, Cringley: Just Say No: David Harrison wants to replace your Internet

Robert Cringley talks about a plan from David Harrison to operate a competing DNS server to ICANN’s “official” Internet server – something Harrison calls Inet. This would effectively block out spammers, spoofers, squatters, and other nefarious s*ers by denying them domains on the new network, but would be backwards compatible with the ICANN DNS servers. It could be as simple as pressing a button on a Web browser to switch the old network to the new one, and vice versa.

Inet would operate its own DNS system parallel to the one run by ICANN. That's not really such a big deal, you know. Certainly a different DNS with different rules would not be hard to build from a technical or even a financial standpoint, and it could exist on the current network right alongside the current DNS system. The big question is why people would use it. They wouldn't at first, because without traffic and participating servers such a DNS would be useless, and that's why David proposes an Inet DNS filter as a crossover between the old/evil system and the new/good one…

What David Harrison is proposing isn't all that different from what happens when a nation replaces its currency, eliminating overnight through the substitution of new paper the counterfeiting, theft, and improper distribution of wealth that had come to characterize the previous currency. If you do it once you'll have to do it again, of course, but even if the changes happen only every decade, wouldn't it be worth it?

With a promise to “prohibit child pornography, phishing, fraudulent commercial services, spam, denial of service attacks, and zombie networks,” we’re wondering if this idea is quite the panacea as its proponent claims – nevertheless, I can see the value in creating an own, private DNS server specifically for business communication, and if you can have your Inet cake and ICANN too, the idea is worth consideration.

GigaOM: Desktop Apps, Reborn as Hybrids

Om Malik talks about the idea of “desktop services” – hybrid web applications that use client software to work offline, then connect to the Web for data storage and retrieval.

Zimbra is not the only one that has realized that many of us want our emails while sitting in a plane, or where Wi-Fi connections are not as commonplace as the Starbucks logo might suggest. Even Salesforce.com has an offline edition of its offerings. Some independent developers are attempting to build offline versions of GMail and some of 37Signals’ web applications.

I’ve often felt that there were certain offline apps that were simply better than online apps, ever since the days that my choices for accessing my college e-mail were Eudora Light or telnetting into the server and using PINE.

More importantly, having an offline app for SAAS means that SAAS sessions will become more popular (increasing the need for bandwidth) but may not require downloading the UI every time you connect (decreasing the need for bandwidth.) In the end, I’d rather download a single e-mail than download that e-mail, the AJAX code to format that e-mail, the graphics used to make the service look good, etc.

We’re having a Webinar today at 10:00 a.m. PST / 12:00 p.m. CST / 1:00 p.m. EST on “Evaluating WAN Optimization Options and Quantifying the Results.”

It’s co-hosted by Cisco and talks about what WAN optimization solution is best for your organization, and how to prove the effects of WAN optimization on application performance. Speaking will be Vijay Iyer, Product Manager for Cisco Wide Area Application Services (WAAS) and Dr. Steve Fulton, Senior Director of Strategic Alliances at NetQoS.

We’ve set up this area on the blog to get attendees questions before the Webinar so that we can get to them in case the presentation runs long. Please enter your questions below. We moderate comments to keep out spam, so there may be a delay before seeing your question appear. If we don’t get to your question during the Webinar, we hope to produce a Q&A on this blog to answer your question.

(Edit: Thanks for participating in the Webinar. We're working with Vijay Iyer to answer the questions and should have the answers up sometime next week.)

By Beth McAfee

I spent my vacation visiting South by Southwest (SxSW) here in downtown Austin.

Besides the secrets to the success of tech stars everywhere, here are some interesting things I learned at the Interactive side of SxSW, and how they apply to network performance:

• People are considering putting broadband municipal internet over broadcast television channels. Bruce Sterling talked about this possibility in his annual rant. He mentioned a scheme developed by Reed Hundt, former head of the FCC, to sell 700MHz spectrum to emergency service providers and possibly for use as municipal broadband internet. I’m not sure whether this is a real possibility or just a random thought in a rambling, yet entertaining discussion, but it is an interesting idea nonetheless.

  Network Performance Application: Municipal Broadband is also interesting for enterprise developers, who could then cheaply route non-critical Internet traffic over the Muni network, and remove that traffic from the WAN.

(Continued...)

Continue reading "Notes on SXSW" »

By Brian Boyko

From the latest XKCD:

Where did those people in the distant future get the idea that bloggers wore red capes, goggles and blog from high altitude balloons?

The wind would blow papers right off my desk.

That’s why I blog from my secret volcano lair.

I’m hoping that other bloggers out there join me in helping set the record straight about this.

Mil-Embedded.com: Raytheon develops world's first polymorphic computer

Raytheon developed a multi-use "polymorphic" computer whose architecture can change on the fly, reducing the number of different processors for computing systems.

"Typically, a chip is optimally designed either for front-end signal processing or back-end control and data processing," explained Nick Uros, vice president for the Advanced Concepts and Technology group of Raytheon Space and Airborne Systems. "The MONARCH micro-architecture is unique in its ability to reconfigure itself to optimize processing on the fly. MONARCH provides exceptional compute capacity and highly flexible data bandwidth capability with beyond state-of-the-art power efficiency, and it's fully programmable."

Of course it'll be a few years (or months) until polymorphic chips find their way into the server room but those will be interesting times until the "dual use" becomes apparent in optimizing server functions and routing. It continually seems that the only place where human beings can't find new technologies to reduce delay is the speed of light.

MONARCH's polymorphic capability and super efficiency enable the development of DoD systems that need very small size, low power, and in some cases radiation tolerance for such purposes as global positioning systems, airborne and space radar and video processing systems.

Or routers.

Slashdot Commentary on "Dungeons And Dragons and IT"

It's worth it just for the in-jokes of the D&D/IT crowd.

The problem is usually that most companies don't hire any more D&D players than it takes to just barely put fires out. You wouldn't be putting out fires all the time if your employer would hire more wizards, although wizardry doesn't come cheap.

You can get four or five wizards for the price of one, but the catch is, the wizards come with the curse that Rutger Hauer and his girlfriend Michelle Pfeiffer had in that movie Ladyhawke. He was a wolf at night and his girlfriend Michelle Pfeiffer turned into a hawk during the day. A simple email conversation would have taken them days and days!

Just like working with overseas teams. Except neither of us look like Michelle Pfeiffer OR Rutger Hauer.

And this one.

It's not a werewolf, it a guy with a "unix beard(tm)".

The way to check is that you say "say, what's your opinion on packaging systems?" If they growl and try rip your throat out, use the silver bullets. If they start to tell their grand view of how packaging should work, use regular ammo.

If ever a werewolf were to evolve that has a fur pattern that looks the same as a short sleeved shirt with pocket protector, the human race is doomed.

While we're on the topic of IT humor,

Novell "Get a Mac Spoof" #2

This is a series of Novell-backed parodies of the Mac vs. PC ads which have an attractive woman playing the part of Sus(i)e Linux. And I think Novell is hitting that target market. I personally have a PC which dual boots Windows and Linux, and just bought a Mac Mini yesterday (for an article I'm working on) and let me tell you, I'd trade them all for a girlfriend.

By Brian Boyko

“O, Canada,” I sang, “Our home and native land. True patriot love. And all thy son’s com—”

“Brian,” interrupted Andrea in the next office over. “What are you doing?”

The truth is I was working on this editorial, which is ostensibly about VoIP and traffic and networks and data analysis and all that stuff but it begins with me making a fool out of myself singing the (likely misremembered) lyrics of the Canadian national anthem.

So bear with me a bit.

(Continued...)

Continue reading "'I Swear, There is a Good Reason I'm Singing at Work'" »

With quick commentary notes from NetQoS CTO, Dr. Cathy Fulton

Ars Technica: Adobe launches Apollo, its web application runtime for the desktop

Now this is seriously cool – Adobe is developing a runtime library that allows Inter-and-intranet enabled runtime apps to work offline as well as online; allowing Web developers to access, edit, and store information even when the network is down or slow.

Granted, no one likes it when the network is down, but sometimes it does happen. This way, people can still use the Web apps and just have the information be updated when the network comes back up. Additionally, you could theoretically use this to schedule SAAS apps to run at a certain time, freeing up some network bandwidth for peak periods.

Downey then pulled the network cable and showed that new auctions could be created even when offline, and the application would automatically upload them to eBay when the computer was next connected. He also demonstrated how the application could create an Excel spreadsheet based on online data and save it to the desktop.

Adobe is giving the runtime away for free, and it can be distributed with each application or downloaded separately from Adobe's web site. The openness extends to third-party applications as well—users can select "View Source" from a menu to see the underlying code behind any Apollo application.

This is especially cool for people who like to use network apps in Wi-Fi enabled coffeeshops, which often have unpredictable downtimes or crowded pipes. Soon, you’ll be able to use Apollo at Starbucks.

(I still haven’t seen all of Season 2.5 yet! Nobody tell me anything!)

Dr. Fulton: It’s certainly cool to have an app work offline and online seamlessly as the network performance degrades. I wish more developers, more companies, would get excited about ensuring their applications worked well online under different network scenarios, however.

Network World: Stanford Researchers Scheming to Rebuild Internet from Scratch.

The “Clean Slate Design for the Internet” is a Stanford University event where researchers will, essentially, explain how the Internet is good, but it could be better if we just started over.

The researchers already have projects underway to support their effort: Flow-level models for the future Internet; clean slate approach to wireless spectrum usage; fast dynamic optical light paths for the Internet core; and a clean slate approach to enterprise network security (Ethane).

Quite frankly, I’m wary of anything replacing the Internet, in this political and economic climate. The Internet developed under the radar of everyone except the geeks; a “clean slate” would invite so many special interest groups, corporations, and national governments to bring in input that the entire operation may be politically, if not technically unfeasible. However, there’s nothing wrong with hearing out the ideas.

Dr. Fulton: This is primarily a “fishing for funding” academic marketing exercise. But a few gems might be discovered that actually will be of practical future use. And that fundamentally is what research is about.

InfoWorld: The Business Case for Open Source Software

This is an editorial espousing the virtues of OSS software solutions for the “free as in speech” approach, pointing out that proprietary software sometimes has onerous licensing restrictions. It’s an interesting read if you haven’t considered it – but we particularly like this one line in the opening paragraph.

“What we really care about (or should care about) is making a sound business decision regarding software.”

We agree with him – mostly – but we also believe that sometimes the open source solution isn’t always the best solution. By all means, software needs to be able to survive the company that spawns it and licensing should be enabling, not disabling. But sometimes the cost of paying for software is worth it – the classic case is GIMP vs. Photoshop. I use both, I prefer Photoshop, and think it’s worth the money. If GIMP ever became as feature-rich as Photoshop, I’d probably use that instead.

Dr. Fulton: The hidden cost of software is in the additional personnel time required to retrieve/maintain the desired solution. I’ve found some open source solutions to be superior (for my uses) to commercial solutions – I use them even when the commercial option is freely available (meaning, the company already purchased it). For large-scale tasks, however, open source solutions usually are not viable. Their hidden cost outweighs their flexibility.

UBCD4Win.com: How to Build the Ultimate Boot CD for Windows

Essentially, this allows you to create a “LiveCD” version of Windows for troubleshooting Windows systems. I often use LiveCDs – Ubuntu mainly – to troubleshoot family computers that won’t boot, but now it’s great to have an option to use Windows. I’m sure many of you already know how to do this but if you don’t, I’m sure you can come up with situations where knowing it would be useful.