Weird Windows XP TCP Behavior

Sam Jansen of Wand Network Research Group in New Zealand recently complained of “weirdness” with Windows XP (SP2) and TCP when doing a TCP test of two systems connected over a link (very similar to those demos at InterProphet). After all, what could be simpler than a couple of cans and a string, right?

No, it isn’t simple. He finds that Windows is sending data outside of the receiver’s advertised window, as well as sending “weird sized” packets in what is supposed to be a “simple bulk data transfer (often sending packets with a lot less than an MSS worth of data)”. What’s going on here?

Poor Mr. Jansen is not losing his mind – what he is seeing is real and Microsoft is cheating. We saw lots of little cheats like this when we were testing on Windows for SiliconTCP and EtherSAN back in 1998 to the present. In sum, Microsoft does this because they think they “get ahead” with a technique called “oversending”. It thrives because TCP congestion control algorithms are all pessimistic with the send budget. It doesn’t always work, like any cheat, but I guess it makes them feel good.

One Billion Hosts and Nothing’s On!

According to Mookie Tenembaum’s latest piece, “The real secret of achieving long “real time spent” is in providing the consumer with something MORE than just the ad message”.

What’s this mean? Content sites can’t just put an ad in front of people without some substantial content attached. It’s a fair exchange – I watch a 15 second car ad on Cnet, for example, for a two minute discussion of the latest digital camera trends by a respected reviewer. Actually, I’m not tuning in to watch the ad, but I am watching it anyway. By gum, isn’t that just like TV?

So how much can customers tolerate – the average length of streaming video is 2.5 minutes with one (or several) ad, total running time 10 seconds max.

So why are publishers complaining about bleeding ink? They’ve already got the editors, writers, and customers reading their magazines and newspapers – all they have to do is put them in short interesting videos talking about what they do!

Of course, there’s that bugbear of “short interesting videos”. Last time I talked with a publisher, I heard all about “that video guy” who never delivered, never finished on deadline, and never made anything interesting. I guess that’s why I work for ExecProducer and use MinutePitch for polished videos in minutes. Because I can’t wait around for “that video guy” – time is money.

Why Women Don’t Like IT?

Ed Frauenheim of cnet put together an article on why women have trouble with IT. So I start talking, and before I know it I’m sitting in CNET’s letters section next to one of RMS’s rants. Good show!

It’s a good article, and I’m pleased to see this subject is starting to be discussed more. When I wrote about this serious loss of women in computing (SF Chronicle, Sept2003, Paving the Way for Systers reflecting on the passing of Anita Borg and the impact of women on technology, I found a real dearth of discussion of this issue in the mainstream and technical press – it was viewed solely as a “woman’s issue” relegated to the margins. Yet I received a great deal of email, both from women and men who have lived with this problem when the article appeared – much more than usually is received. And what they shared with quite striking.

You Can Have it All – Unless You Want to Retire

Sue Hutchinson wrote a nice article that discusses the closing of the gender gap in longevity. But I’m hopefull for a followup. Seems that there’s serious talk in the social security reform set of taking into account women’s “longer” lifespans by reducing benefits (e.g. Meet the Press last month) right when we don’t seem to have that edge anymore:

“MR. RUSSERT: Do you think Congress, Mr. Chairman, would accept any formula that said that people would be treated differently because of their gender or their race?
REP. THOMAS: If we discuss it and the will is not to do it, fine. At least we discussed it. To simply raise the age and find out that you’ve got gender, race and occupational problems later, I would not be doing the kind of service that I think I have to do. You and I have been around quite a while. We went through the ’80s. We went into the ’90s. And now we’re in the 21st century. We saw the choices that were made in the past. We went to the well over and over again with the same old solutions which really aren’t solutions. We’ve reached the point where we have to fundamentally examine it in my opinion. The president has given us that opportunity. We ought to take it.”

So what if there is no “longer” lifespan for women within the next 30 years if these factors Sue described continue? The reset for social security is then based on an outdated premise? Coupled with lower lifetime earnings by women, looks like the reforms in social security could create a new subclass of permanently improverished women.

The Limits to Internet Media is the Content

Fascinating little keynote at Imedia Summit by Lincoln Millstein, COO of New York Times Digital on the future of media.

Mr. Millstein sees “Big Iron Publishing” – the presses and paper and trucks and newsroom – as being nonscalable and noncompetitive compared to Internet media. Also, inventory of content is the burning limiting factor to encourage more interaction and stickiness with a site’s audience.

What Makes a Great Venture? It’s the team, execution, and trust…

John Doerr spoke with his usual inspirational passion at the Stanford Entreprenuerial “Thought Leaders” seminar. He asked “What makes a great venture?”. Well, I’ve been around the startup block a few times or more, so that’s easy – team and application. “What makes a great company?” How about focus and execution. And “What makes a great group?” Piece of cake – interdependence and communications. And most importantly – trust.

I’ve heard most of this stuff time and time again, so I was jazzed he wanted to skip a lot of the canned presentation and move on to questions. So my question that got written under the biz catagory was:
“WSJ reported last week that ad revenues are finally moving substantially to Internet / online from print. Does this mean, as some claim, we’re in a “second Internet bubble”, or do you think it’s just the promise of the first Internet rush finally realized – in other words, it just took longer than expected.”

While a lot of others were of the “gee, what’s hot to invest in” and “how about my pet venture” variety (and dispatched quickly), mine made him pause a while and think about it. Can you see why?

Open Source Software Patents an Oxymoron?

“Recent events have drawn attention to the intersecting realms of patents and open-source software. IBM has donated 500 patents for use in open-source software, Sun plans to liberate 1,600 for use with its open-source Solaris operating system, and a Hewlett-Packard executive believed in 2002 that Microsoft planned to attack rival open-source projects with its patents,” according to Stephen Shankland of Cnet. That should be good, right?

Linus Torvalds, Brian Behlendorf and Mitch Kapor don’t think so. Even with donations of patents and white knights like IBM to protect Linux, other companies like Microsoft could launch a patent attack that could devastate open source. “We have to be concerned about…the use of patent WMDs. That will be the last stand of Microsoft,” said Kapor.

Aside from the melodrama, is there a case for the granting of software patents? Back in the 1980’s, when Unix was in its heyday, software patents were not filed because software was considered a collection of algorithms (not patentable) that was “expressed” in a unique form similar to that of a book or music. Hence, software was protected under copyright and trade secret status.

It seems inevitable that the success of open source spawned the boom in software patent filings, since trade secrets can no longer be maintained in open source and copyrights are routinely ignored. Patents are published but provide the grantee with exclusive ownership of the process for close to two decades. This made sense when hardware manufacturing processes and inventions might take many years to finance and develop, but not in the case of an intangible like software.

I expect there will be more filings – not less – for software patents. In operating systems alone I routinely review patents that are mere windowdressed versions of basic virtual memory, filesystem, stack, and driver functions that predate my involvement at Berkeley. The temptation to file by big entities (remember – “PatentLand” is pay-to-play) is just too great.

Perhaps a very short-term protection status for software patents (like three years) with a separate fast-track review process might be a far more reasonable solution to the current dilemma.