Practical Technology

for practical people.

March 23, 2005
by sjvn01
0 comments

Software Patents and Mutually Assured Destruction

The Cold War policy of MAD was that things could never get too ugly between the United States and the Soviet Union because if one of us went too far, the other could bomb it into the stone age … and vice-versa. That may be where we’re
Have I mentioned that I hate software patents?

Why, yes, I have. Many times.

Im also a realist. Maybe someday, PUBPAT (the Public Patent Foundation) and other patent-reform groups such as the Electronic Freedom Foundations Patent-Busting Project will succeed in getting rid of software patents, but Im not holding my breath.

Even with Microsoft backing some software patent reform, amazing but true, I just cant see significant changes to U.S. patent law happening anytime soon.

And lest we forget, Microsoft isnt exactly the most trustworthy company in the world when it comes to patents.

Just this week, Microsoft announced that a new, integrated IPv4/IPv6 network stack would be in its upcoming Longhorn operating system. At the same time, lawyers for the Public Patent Foundation and the Software Freedom Law Center are concerned over a patent about automatic generation of IP addresses to facilitate simple network connections.

If my network-savvy readers say that sounds a bit like some of the features of IPv6 and DHCP (Dynamic Host Configuration Protocol), well yes it does, doesnt it.

Now, IPv6 isnt that important in North America at this time, but sooner or later, everyone on a TCP/IP network connected with the Internet is going to have to deploy it. Some people, with good reason I think, are concerned that Microsoft might use that patent to try to place a tariff on all programs that access the next generation of the Internet.

This is only part of what many see as a pattern of Microsoft getting ready to retroactively claim IP (intellectual property) rights over many of the Internets basic protocols.

Maybe they are, but Darl McBride, the CEO of The SCO Group, made me think that that may not be the case. We were talking about patents on Tuesday—no, SCO is not getting ready with a surprise patent lawsuit—and he said he didnt think Microsoft would dare use its patents in such a broad way.

His logic was that the technology world is a lot like the world situation of the 50s through the 80s, when the Soviet Union and the United States never went to outright war with each other because of the cynical but all too practical notion of MAD (mutually assured destruction).

In a nutshell, the policy of MAD was that things could never get too ugly between the United States and the Soviet Union because if one of us went too far, the other could bomb it into the stone age … and vice-versa.

How that applies to todays computing world is that if Microsoft started really throwing its patent weight around, IBM or Novell could retaliate in kind. Thus, if any one company tried to really strangle a large part of the market with an overly aggressive patent enforcement, they would be blasted by other companies with large patent portfolios. The end result would be that all of the companies involved would be locked into a software development doomsday, where nothing could be developed.

This isnt just idle speculation. After talking with McBride, I contacted several other people. Both a senior Novell executive and a prominent open-source attorney told me that if Microsoft ever tries to push too hard with its patents, other companies are more than ready to counter-attack with their own patent portfolios.

Maybe they are, but Darl McBride, the CEO of The SCO Group, made me think that that may not be the case. We were talking about patents on Tuesday—no, SCO is not getting ready with a surprise patent lawsuit—and he said he didnt think Microsoft would dare use its patents in such a broad way.

His logic was that the technology world is a lot like the world situation of the 50s through the 80s, when the Soviet Union and the United States never went to outright war with each other because of the cynical but all too practical notion of MAD (mutually assured destruction).

Lets take this analogy a little further. The peaceful coexistence based on MAD between the Soviets and the Americans meant that we didnt see World War III. We did, however, see lots of brushfire wars in Afghanistan, Angola and Vietnam.

Thus, just because the big companies may never go too far with their patents with each other and worldwide technologies such as the Internet doesnt mean that smaller companies or open-source developers wont be attacked. They will be.

Of course, as the good people at Black Duck Software will tell you, a patent holder doesnt actually have to sue you for a violation. Just the mere threat is enough to stop a company from developing or marketing a program if it doesnt have the legal protection or deep pockets needed to fight a patent battle in the courts.

So, while I now think that were unlikely to see any truly broad attempts to abuse software patents, I still believe that software patents are bad in general and that, in specific, small companies and open-source developers are still vulnerable to the bullying of software-patent superpowers.

A version of this story first appeared in eWEEK.

March 21, 2005
by sjvn01
0 comments

Look Out, Windows: Its the Next SuSE Linux Desktop

The company’s next-generation enterprise desktop, Novell Linux Desktop, won’t be out until 2006, but Novell’s SuSE Linux Professional 9.3 desktop will have many of its features and will come out in April.

At its annual BrainShare tradeshow Monday, Novell showed off the next edition of its enthusiast Linux desktop, SuSE Linux Professional 9.3, to a few partners here.

The SuSE Linux Professional desktop will ship in April. It comes with all the usual updates one would expect from a modern, full-featured Linux desktop: the Firefox 1.01 Web browser, the Novell Evolution 2.0 groupware client, the 2.6.11 Linux kernel and your choice of a KDE 3.4 or GNOME 2.10 interface.

But theres more. In addition, this distribution will ship with OpenOffice.org 2.0, the latest version of the popular desktop application suite; VOIP (voice over IP) Linphone; and iPod compatibility.

SuSE Linux Professional, though, will not just be a collection of cutting-edge Linux and open-source technologies. SuSE is adding its own new contributions to the Linux desktop.

This distribution is also the forerunner for Novells next enterprise Linux desktop: NLD (Novell Linux Desktop) 2.

In fact, Nat Friedman, vice president of Linux desktop engineering at Novell Inc. said, “We are getting ahead of Windows for the first time.”

While the new Professional hasnt been shown to the public yet, those bedazzled attendees who have managed to see it think Friedman may be right.

The most exciting development, according to several observers, was Beagle. This is a new integrated desktop metadata search program that indexes a users hard drive content in a way thats similar to Google Desktop 2 or Apples Spotlight. One observer commented, “Beagle is like Desktop on steroids.”

Click here to read more about Novells annual trade show, BrainShare.

Beagle, which uses the Mozilla Gecko rendering engine for its display, is more than just an attractive, interactive search tool. It can also be used in Mono-based C+ programs, as Friedman demonstrated during an interview.

A demonstration of some Beagles features is available online.

Friedman also said that with Beagle in place, it would finally be possible to create a program he calls “Dashboard.”

Dashboards goal is to automatically show a user useful files based on his or her current activities. For example, if you were writing an e-mail about your plans for the coming weekend, Dashboard would automatically display your schedule for that weekend.

While Dashboard wont make it into 9.3, Friedman has hopes that it will make it into the next version of Professional, and possibly into NLD 2.

The new Professional also comes with another program thats reminiscent of a Google software offering. In this case, its Googles free photo-editing, organizing and Web publishing program Picasa 2. Novells answer to this Windows-only program is the F-Spot photo-management tool.

Both these innovative Novell-sponsored programs are licensed under the GPL. Beyond the feature set, Professional will also include Xen 2.05 virtualization software. This will enable users to run multiple instances of Linux simultaneously.

While this may seem more like a parlor trick than something useful on a desktop machine, Chris Schlaeger, Novells vice president of engineering for Linux, said, “It will work on Professional, but were not going to say its ready for enterprise use, because its not.”

Looking ahead, though, Schlaeger said, “Once Xen has been mastered on the Professional platform, its virtualization capabilities will become part of our data-center offering, where the processor virtualization is extremely important.”

And looking ahead for Professional, Ed Anderson, Novells vice president of Linux and platform services, hinted that Novell may be thinking about opening up its development path. “Weve been mulling over creating a community-based Linux like Red Hats Fedora. There are many things we like about the Fedora Project, but there are some we dont.”

A version of this story was first published in eWEEK.

February 11, 2005
by sjvn01
0 comments

What Comes the Day After SCO Dies?

Some folks I know over at SCO are peeved that no one sees that they won a victory in the latest clash in the SCO vs. IBM saga.

One such objector said, “If the whole world took everything that some people said as the gospel truth, then no, this wouldnt look very good for SCO. How theyre able to take the judges denial of all of IBMs Preliminary Summary Judgments and say that this is an awful day for SCO is beyond me though. I wouldnt say that it was a win for either side, but it also wasnt a one-sided loss for SCO as some would have the world believe.”

Yes, looked at from that point of view, it was a “victory.” But theres a phrase for this kind: “Pyrrhic victory.”

Pyrrhus, for those who were snoozing during history class, was a king of Epirus in Greece who fought with the Romans. He won several victories, but after the last one at Asculum, seeing how many men he had lost, he said, “One more victory like this will be the end of me.”

Exactly.

When the judge doesnt close the case against you, but writes, “It is astonishing that SCO has not offered any competent evidence to create a disputed fact regarding whether IBM has infringed SCOs alleged copyrights through IBMs Linux activities”; and notes that there is a “vast disparity between SCOs public accusations and its actual evidence—or complete lack thereof,” its not much of a victory.

So, it is then that I started thinking about what might happen after SCO loses.

All SCOs copyright claims over Unix are open to debate, even if you discount Novells claims. The courts will make the final call, but I dont expect SCO to end up with much of anything.

Still, SCO has some original work in OpenServer and UnixWare Unix variants, and some of it, like its SMP (symmetric multiprocessing) and OpenServers incredible stability, is darn good. In addition, there are still quite a few resellers and their customers who rely on their products.

With Canopy, SCOs parent company, in disarray with its former leadership fighting with its founders family, I have a bad feeling that SCO wont survive at all.

Prior to the Canopy fight, and with its legal costs capped, I had assumed that the post-defeat SCO would slowly decline with support for its loyal customers as they gradually migrated away to Linux or Solaris on x86.

Now, now Im not sure it will be able to do that. SCO could be here one day and gone the next, depending on what the courts decide.

What I think will happen is that someone—Novell? Sun?? IBM!! Microsoft!!?—will buy what Unix IP SCO is left with. As for its products… OpenServer and UnixWares market share was in decline long before Linux started taking over the Unix world.

Still, someone, I think, will take them up. Theres still money—as many a reseller can attest—in supporting these systems. But it wont be much money, and you certainly arent going to see your income grow as time goes by. I suspect that one of the bigger SCO resellers will end up with SCOs two Unix operating systems.

Were HP not in an uproar, Id think that company might make a good buyer for these operating systems. Its not well known but HP, via Compaq, has a long relationship with SCO and its Unix products. With Carly out, though, I think thats out of the question.

Novell, which had strong ties with SCO, and certainly knows about the problem of moving customers from older operating systems to Linux, might also make a good buyer for SCOs operating systems. And, since Novell is in the area, it would also make a natural employer for displaced SCO employees.

One thing though is for certain. Its time to start thinking about a post-SCO world. It may still be years away, for the mills of justice grind very slowly, but the day will come.

Oh, and as for Linux? Hello, Mr. Gates, your worst nightmare is coming true. After burying all your opposition, youve got a real operating system competitor on your hands again.

A version of this story first appeared in eWEEK.

February 3, 2005
by sjvn01
0 comments

Sun, Open Source and Star Trek

One of classic Star Treks best episodes was “Mirror, Mirror” For those of you who have somehow managed not to see it, in this episode Kirk, McCoy, Uhuru and Scotty are caught by an ion storm in mid-beam-up and end up in an alternate universe Enterprise, crewed by evil versions of their comrades.

I dont know about you, but when I look at Sun and its recent open-source moves, I start seeing an open-source “Mirror, Mirror.”

For example, a few weeks ago, Sun announced that it was releasing patents to the open-source community.

Now, as someone who hates software patents, I like the idea of anyone—Novell, IBM, Sun, whoever—opening their software patents up to the community. But, questions began to spring up about exactly how open Sun was being with its 1,600 patents.

Dan Ravicher, the Public Patent Foundations executive director opened up this can of worms by writing in an open letter to Sun, “In Suns announcement, they make sweeping statements about how the open-source community will immediately gain access to 1,600 active Sun patents for operating systems, but the legal nitty-gritty behind the announcement shows that Sun has retained the right to aim its entire patent portfolio at GNU/Linux or any other free and open-source operating system, except, of course, for their soon-to-be-released version of Solaris.”

Bruce Perens, one of the founders of the Open Source Initiative, put it more bluntly on Slashdot: “They can sue Linux developers over those patents. They can sue their own Open Source partners.”

Then, earlier this week, Suns head of Solaris marketing, Tom Goguen, tried to clear up the situation by saying, “Clearly we have no intention of suing open-source developers.” But, that, “We havent put together a fancy pledge on our Web site” to that effect.

Well, if you trust Sun, that might be good enough for you. But, if theres one thing Ive found over the years, its that Sun is constantly changing its stance, and it has happened yet again.

On Tuesday, during the question-and-answer session at the keynote speech of Suns Network Computing meeting at the companys Santa Clara, Calif. campus, COO Jonathan Schwartz said that such opposition came from people who “believe that everything must be GPL. The open-source community is far larger than just the GPL community.”

Fair enough, but then Scott McNealy spelled out that, “Sun has an obligation to its shareholders to leverage and protect its intellectual property. We are granting [access to our intellectual property] to people who are responsible and who are signed-up licensees of the CDDL.”

Argh! There you go, folks. Sun just doesnt get it.

The whole point of open-source is that by sharing IP, everyone benefits, but you do have to really share it, and Sun doesnt get this.

Larry Rosen, a partner in the technology law firm Rosenlaw & Einschlag and author of “Open-Source Licensing: Software Freedom and Intellectual Property Law” has another worry about the CDDL.

“My biggest concern about the proliferation of reciprocal license such as the CDDL is that we end up not with one commons of free software but multiple islands of it that cant be interchanged for creating derivative works,” Rosen said. “We get some of the benefits of the open-source paradigm but—as the Apache foundation is so fond of reminding us—reciprocal licenses prevent free software from being available to absolutely everyone for modification and reuse.”

Exactly.

Now some people at Sun, like Suns chief technology evangelist, Simon Phipps, think that Im anti-Sun. In his blog entry of Jan. 26, Phipps wrote, “As usual Steven Vaughan-Nichols of eWEEK is the spokesman for the competition and the leader of the (many) journalistic naysayers who cant just bring themselves to admit Sun might have done a good thing.”

Actually, over the years, Ive liked Sun. I still like Solaris—I just dont think it will ever be open-source—but what I dont like is Suns constant game playing.

Phipps then went on, “Just to show how far over the edge SJVN and his usual suspects are, take a look at Groklaw, where Pamela Jones has written a long, thoughtful and balanced analysis of the news. It is the opposite of the trade press shallowness.”

I wonder if Phipps feels the same now that Jones has listened to the same comments that I have and what he thinks of her latest: “So, about that CDDL. Watch out. Thats what Id say. Use it only if you trust implicitly in Sun. And if you do, Id certainly like to know why. The community needs to watch this company like a hawk, in my view, after what I saw today. They are not yet full members of the Open Source community, to phrase it as positively as I can.”

You can say that again.

Listen, Simon, Scott, Jonathan, forget about calling what youre doing open-source. No one says you have to be open-source. Well, OK, so you need to do something about the fact that Linux is eating Solaris market share lunch, but I think youve already got that with your N1 grid model.

Selling it isnt going to be easy, because grid is a new idea, but, from that same speech, it sounds to me like youre finally getting the grid message across. Heck, maybe you can make it profitable at a buck an hour.

And, you can do that without simply calling what youre doing open-source, because, the bottom line is, its not. And, if you keep trying to have it both ways, Sun may end up like the latest version of Star Trek, Enterprise: cancelled.

A version of this story first appeared in eWEEK.

January 7, 2005
by sjvn01
0 comments

MIMOing Your Way to Faster Wi-Fi

Wi-Fi vendors Netgear Inc. and Linksys, a division of Cisco Systems Inc., have both announced this week that they will be releasing Wi-Fi products that use MIMO (multiple input, multiple output) antenna technology to improve their 802.11g products’ speed, range and reliability.

Linksys claimed that its new MIMO-enhanced 802.11g devices would enable users to have three times the range, and network speeds of up to eight times that of conventional 802.11g. This, if accurate, would give Linksys MIMO devices throughputs in the range of 400-Mbps.

More realistically, most vendors predict that users will see at least two to four times faster throughput and a doubling of range. Perhaps more importantly, MIMO Wi-Fi devices should deal better with structural issues that lead to dead spots in Wi-Fi coverage.

Best of all, while neither Netgear nor Linksys has announced prices yet, it is expected that MIMO adaptors and access points will cost little more than their existing 802.11 counterparts.

How can they do it? MIMO works by taking one of radio communication’s oldest problems, multipath, and turning it into a solution.

Multipath is what happens when signals bounce off objects or structures and take multiple paths to the receiver. If you listen to your car radio, you run into the multipath problem every day. For example, if your favorite radio station fades out every day in a certain location, you’re hearing an example of what’s called multipath fading or “Rayleigh” fading.

What’s happening is that your antenna is receiving both the transmitter’s main signal and its reflections. When these signals arrive out of phase with each other, they cancel each other out, and your morning traffic report fades out.

But starting in the 1990s, a pair of Stanford University researchers showed that you could use each reflection, each multipath route, as a separate channel.

The engineering problem was that in order to make use of multipath this way, you need multiple antennas. However, it turns out that these antennas can be very close together- close enough to fit on a Wi-Fi card.

So it is that MIMO devices actually transmit and receive multiple data streams over their multiple antennas. These streams are then bonded together on the Wi-Fi device to create a higher-speed wireless connection.

If this sounds familiar, it should. Bonding was a trick we often used with ISDN (Integrated Services Digital Network) routers, and in modem devices that combined several 56K modem connections into a single, faster connection.

More recently, Wi-Fi chip maker Atheros Communications Inc. this week has used bonding with 802.11h to make Wi-Fi devices that could double the theoretical speed of 802.11g from 54Mbps to 108Mbps. Wi-Fi devices labeled “Super G” use this technology.

Unfortunately, these accomplish this speed increase by bonding together two or more of the 802.11g standard channels. Therefore, if you have multiple 802.11g networks, they can interfere with each other. This can result in a slowdown in the conventional 802.11g network.

MIMO avoids this problem by not bonding together 802.11 channels. Instead of sending one data stream down one channel and another stream down another channel, MIMO simultaneously transmits multiple data streams over the same channel.

As you might guess, this does cause signal interference. But MIMO receivers use algorithms to pull out the proper data streams and bond them in real time, resulting in a much faster throughput with longer range than conventional 802.11 technologies.

MIMO also uses SDM (Spatial Division Multiplexing). SDM multiplexes the multiple data streams, one per antenna, to transfer data simultaneously in each channel of bandwidth. This also results in faster traffic while still using any of the 802.11 networking protocols: 802.11a, 802.11b or 802.11g.

Airgo Networks Inc. was the first company to release a commercial version of MIMO. The groundbreaking firm’s “True MIMO” AGN100 Baseband/MAC (Media Access Control) processor and AGN100RF transceiver are being adopted by such Wi-Fi equipment vendors as Belkin Corp., Linksys and SOHOware Inc.

Airgo, however, is far from the only chip OEM to be producing MIMO-capable chip sets. For example, Netgear is using Video54‘s RangeMax MIMO technology. Intel is also working on MIMO technology for its next-generation Centrino chips.

Of course, it would be too much to expect these varying MIMO implementations to be compatible with each other. Each of them, however, is backwards-compatible with existing Wi-Fi equipment. In addition, to that extent, pieces of MIMO equipment will be compatible with each other. So, for example, while you wouldn’t get 100Mbps from a combination of Belkin MIMO-enhanced 802.11g access points and Netgear MIMO-enabled Wi-Fi cards, you would still get 802.11g’s usual speeds.

In addition, MIMO Wi-Fi cards, even used with non-MIMO access points, should see extended range. That’s because, by their very nature, these cards are better at picking up low-strength signals.

What all of this means for you as a reseller or an integrator is that MIMO technology deserves your attention today. Everyone loves Wi-Fi, especially now that it’s getting more secure with the maturation of 802.11i. At the same time, everyone wants more range and faster throughput.

If at all possible, stick with one vendor, or at least one MIMO chip set, in your deployments. If you don’t, your customers won’t see significant performance gains.

Finally, if you’re hoping that 802.11n, the next high-speed networking update standard, takes care of these incompatibility woes … don’t hold your breath. 802.11n is in the very early stages of development, and we’re already seeing at least two main camps with different ideas on how to achieve a real-world 100Mbps Wi-Fi throughput.

On one side, you have the MIMO camp led by Airgo. They’re grouped together under the name WWiSE (World-Wide Spectrum Efficiency). On the other, you have the TGn (Task Force 802.11n) Sync group. This group, which includes Atheros, Intel, Nokia and Sony, is taking the bigger-is-better approach by using existing wireless technologies over bigger, 40-Mhz channels.

The bottom line? Don’t wait for 802.11n, but do test out the coming MIMO equipment, pick a vendor and start rolling it out. You and your customers will be glad you did.

A version of this story first appeared in Channel Insider.

January 2, 2005
by sjvn01
0 comments

The Triumph and Tragedy of Ray Noorda

Over my years as a computer journalist, I’ve met pretty much everyone who’s anyone in the technology business. Very few of the big names have really impressed me, though. Among those who I’ve thought were really something special have been Steve Jobs of Apple, Linus Torvalds of Linux and, oh yes, Ray Noorda of Novell.

Unlike those others, Noorda wasn’t technically brilliant. He was just a hard-headed, hard-working businessman who made things happen. I admired his work ethic and his willingness to do what was right by his lights.

Sometimes that meant fighting the good fight against Microsoft, which he saw as potentially totally dominating the desktop long before everyone else did. And sometimes that meant making peace.

For example, after Novell bought Unix and USL (Unix Systems Laboratories) from AT&T, rather than continue to fight with BSDI (Berkeley Software Design Inc.) over possible Unix intellectual property rights violations in BSD/OS, an early, commercial BSD Unix, Noorda famously declared that he’d rather compete in the marketplace than in court, and the two sides settled peacefully.

Noorda also took Novell Data Systems and turned it into the LAN company of the late 1980s and early 1990s. Along the way, Noorda also shaped the modern reseller world.

It was Noorda who came up with the three-tiered approach to resellers that so many companies now use. It was Noorda who decided that training and certification were the best ways to make sure that his partners could support Novell’s products. In turn, those early Novell resellers used Novell’s CNE (Certified NetWare Engineer) to market their wares and services so successfully that every major vendor has a certification program for its partners.

If you use any certification to get a job or to sell your services, you owe a debt to Ray Noorda.

Thus, it’s with a heavy heart that I see that Noorda’s investment company, The Canopy Group, parent company of SCO, and many others involved in a civil war. It appears that an old-fashioned power struggle has developed between Noorda’s hand-picked executives, some of his family and some Canopy associates.

I don’t know the details of the fight. No one outside of the Canopy family does.

But I do know that Canopy, which is a privately held venture capital firm, almost certainly has more than $1 billion in assets. It doesn’t take a genius to see how people could fight over it.

It also doesn’t help any that Canopy’s internal structure is a Byzantine complex of, at one time, more than two dozen companies. These companies frequently share resources and officers and invest in each other.

For example, in late 2000, SCO (then Caldera) formed a partnership with EBIZ Enterprises, a Texas-based company in which Canopy held a controlling interest, to create a B2B (business-to-business) marketplace company called PartnerAxis. The company was designed to link Linux solution providers with resellers. PartnerAxis, in turn, was funded by Canopy.

I’m no accountant, but I have followed several of the Canopy-related companies, such as SCO and Novell, for more than a decade … and I can’t make heads or tails of Canopy’s internal structures. If you want to give it a try, the TWkiIWeThey site, with its listings of Canopy officials and companies, is your best place to start.

Novell, I should add, isn’t part of Canopy anymore. Noorda and Canopy largely divested themselves of Novell stock back in 1996.

In part, that was because—oh, the irony of it all—Noorda believed that Linux was the future, while Novell, now under Bob Frankenberg, killed off its internal Linux skunks work project. Soon thereafter, Noorda cut off his ties with Novell and via Canopy bankrolled Caldera Systems, one of the first Linux companies.

Over time, though, Ray’s health and memory have continued to fail. And meanwhile, Caldera, now renamed SCO, has turned against Linux and Novell has now embraced Linux with open arms. Oh the irony!

Were Ray, who is 80 now, in better shape, I’m sure that we would never have seen SCO sue IBM and launch its attack against Linux. Instead, we would have seen Novell acquire SCO, rather than SuSE, to make its Linux play. And we certainly wouldn’t see what promises to be a long, nasty fight for control over Canopy.

So it is that today I sit in my office, and I’m very sorry indeed that Ray’s legacy has come to this.

Still, in the larger sense, his legacy still goes on. Regardless of what happens to the companies he founded, he, just as much as Scott McNealy of Sun and John Chambers of Cisco, helped create the networked world we live in now, And more so than anyone else, he created the modern reseller and integrator world from which so many of us make our living.

A version of this story first appeared in Channel Insider.