LWN.net Weekly Edition for April 24, 2003 [LWN.net]

DARPA Cancels OpenBSD Funding

[This article was contributed by Joe 'Zonker' Brockmeier]

If you follow the news at all, you've probably already heard about the OpenBSD project losing the funding from the U.S. Defense Advanced Research Projects Agency (DARPA). What's less than clear is why the funding has been pulled. In fact, it's quite a test to figure out who's actually responsible for pulling the plug, much less the reason. DARPA is, essentially, just an intermediate agency for the funding, which is passed on to the University of Pennsylvania. The funds themselves come from the Air Force Research Laboratory.

Most speculation has gone to comments made by OpenBSD project leader Theo de Raadt. The comments in question come from an interview in The Globe and Mail, where de Raadt is quoted as saying he's "uncomfortable" about the source of the grant. De Raadt also told the Globe and Mail that, "I try to convince myself that our grant means a half of a cruise missile doesn't get built," which might not sit well with U.S. military types. A few days after the comment appeared in the Globe and Mail, de Raadt was contacted by University of Pennsylvania professor Jonathan Smith. According to de Raadt, Smith objected to the comment, but wouldn't give a specific reason why. The funding was pulled on Thursday of last week.

If that is the reason for the cancellation, it's not the official story from DARPA, in as much as DARPA has or will give an official story. A statement forwarded to LWN by de Raadt, attributed to DARPA spokesperson Jan Walker, claims that the funding is under review.

As a result of the DARPA review of the project, and due to world events and the evolving threat posed by increasingly capable nation-states, the Government [sic] on April 21 advised the University to suspend work on the "security fest" portion of the project.

Walker did not respond to e-mails or phone calls requesting confirmation of this statement or requests to elaborate on or clarify the statement.

The most immediate consequence is that the OpenBSD project has had the rug pulled out from under them with regards to the upcoming hackathon in Canada. 60 OpenBSD developers are scheduled to travel to Canada for the event, almost all of whom have already purchased tickets based on a go-ahead given in January. The hotel was contacted and told to cancel the reservation, despite the fact that an 80% cancellation fee is in effect. According to de Raadt, this amounts to about $24,000 Canadian. De Raadt also reports that the hotel was instructed not to allow anyone to pay the remaining balance to keep the reservation. However, de Raadt said that the hotel has agreed to cut the OpenBSD project a deal for the hackathon, even if they cannot apply the cancellation fee to the bill.

Fernando Pereira, chairman of the Department of Computer and Information Science at the University of Pennsylvania sent this statement to the OpenBSD "misc" mailing list to explain why the cancellation fee cannot be used towards the hotel costs:

When the contracting agency requested that work be stopped on the security fests component of POSSE, the only expenses that they would still allow are documented losses to the conference hotel due to cancellation. Any other use of funds, including use of the cancellation costs in partial support of conference accommodation, would not be an allowable contract expense. Contrary to a widespread misconception, the University of Pennsylvania could not have "allowed" that use of US Government funds. The funds belong to the US Government, not to the University.

Apparently, quite a few people in the OpenBSD community have already sent letters of protest to the University of Pennsylvania, newspapers and other sources. If you'd like to write a letter to complain or comment on the decision to official sources, de Raadt notes that it's helpful to have the contract number. The contract was granted by the Air Force Research Lab, Material Command, and is DARPA contract number F30602-01-2-0537.

With the exception of the hackathon, the loss of funding may not be as dramatic as it sounds. On Monday, de Raadt said that the OpenBSD project had already received about $7,000 in donations, and more was "in the mail." The OpenBSD project has been around for eight years, and has done just fine without the DARPA funding. In addition, the funding was set to run out within four months anyway and de Raadt noted that he works through a Canadian contracting company that should ensure that he receives the rest of his pay for the next four months. The major losers appear to be the University of Pennsylvania grad students who were also receiving money from the grant, as well as the 60 OpenBSD developers who are wondering whether there will be a place for them to stay when they arrive at the hackathon.

Comments (16 posted)

Novell and Linux

Readers of the discussion on LWN.net may have seen comments posted by Kristopher Magnusson, who happens to be the chair of Novell's "Open Source Review Board" and the person responsible for managing the company's relations with the free software community. We had the opportunity to ask Mr. Magnusson a few questions about Novell's plans with regard to Linux; his answers appear below. But first, a couple of other Novell-related items:

Novell has become a gold sponsor of the Linux Professional Institute, and is recommending LPI certification as part of its own certification program.
Jack Messman, Novell's CEO, has sent us a clarification of Novell's view of Linux and the free software community, and an apology for some remarks in an interview that did not come out quite right. " Novell wouldn't be spending the tremendous time, money and resources to make this strategy a reality if we didn't believe in the present and future of Linux. After building and enhancing NetWare for 20 years, this is new territory for us. We simply ask for your patience along the way."

And now, on to the interview.

LWN: In the ComputerWorld interview, CEO Jack Messman said "Linux is an immature operating system right now. It hasn't had somebody like Novell worrying about making it robust, reliable and scalable for very much time. We think we can bring that to the Linux kernel." He has since noted that he could have expressed himself better, and his apologies have been accepted. But the point remains that Novell sees room for improvement in the Linux kernel. The kernel developers agree, of course; otherwise they would be working on something else. Could you explain what improvements Novell would like to see in the Linux kernel?

First, I want to reiterate that Novell believes the Linux kernel is quite mature, robust, reliable and scalable as it is today, or else we wouldn't have decided to use it in NetWare 7. That said, at this point, Novell currently has no definitive plans to improve the kernel, though as Jack indicated we will indirectly enhance it by the services that runs on top. We intend to let the Linux developer community go through its normal development process and use whatever kernel they develop as-is.

Job number one for Novell engineering is to port the services that run on the operating system. Whether customers are running NetWare 7 on the Linux kernel or the NetWare kernel, we want to make sure they have access to the very best services for file, print, storage, directories, messaging, collaboration, resource management, Web development and many others.

LWN: Which of those (if any) does Novell plan to work on (and contribute back) itself?

As I stated, we like the Linux kernel as-is, and have no plans at this point to to develop our own improvements. Novell's focus today is delivering a number of services above the kernel.

LWN: A quick search through the linux-kernel mailing list did not turn up any Novell engineers participating in the discussion - at least, none that identified themselves as such. Does Novell have engineers working on the Linux kernel, and do they plan to participate in the development community?

We do have a team of Linux engineers who have joined the Linux-kernel mailing list and they are reading the Linux-kernel mailing list posts. My understanding is that they are getting a feel for how the discussions take place before they actually participate with questions and so forth--they want to understand the lay of the land before they jump in head-first.

LWN: The recent announcements mention Novell's contributions to various open source projects, including Apache and OpenLDAP. Can you give a quick summary of what some of the more important contributions have been?

Novell has been quietly engaging the open source community for a number of years. For example, our OpenLDAP work has been quietly humming along for four years. And it's not well known that we've thrown our weight solidly behind the "AMP (Apache/MySQL/PHP)" platform that's been so popular on Linux. Because of our AMP work, developers can take AMP code and move it to NetWare 6.5 pretty much unmodified.

Our Apache work is one of our more important contributions. We have a strong relationship with the Apache Software Foundation. In the case of Apache, Novell's lead engineer in charge of porting Apache to NetWare is a member of the Apache Software Foundation, which gives him code check-in privileges as well as some degree of control over the general technical direction of Apache development. Further, Novell has been very conscientious about contributing our improvements to the Apache codebase back to the Apache Foundation.

Novell recently formed a relationship with MySQL AB. We licensed a commercial version of MySQL to ship their database on every NetWare 6.5 CD, and this has been a big hit with our biggest customers. We practice a kind of open source process between our two companies--Novell engineers porting MySQL code make improvements that we contribute back to MySQL AB. These improvements find their way into the GPL version of the database, which benefits everyone who uses the open source version of MySQL.

Novell also has a relationship with the PHP group that's part of the Apache Software Foundation. We ported PHP to NetWare as part of our AMP strategy, and we made a number of improvements to the PHP code that we contributed to that organization.

Beyond AMP, our relationship with OpenLDAP dates back to 1999, when Novell was looking for open source C-based libraries for programmatic access to LDAP directories. We found OpenLDAP's implementation, which needed some work. We decided to pitch in and help; so we completed the work for them and contributed our improvements back to OpenLDAP. Next, we needed a set of Java libraries. OpenLDAP didn't have any, so we wrote our own and contributed them to OpenLDAP outright under their BSD-based license. After four years, we still check in Java library code to OpenLDAP on a weekly basis. Most recently, a few months ago, we contributed to OpenLDAP a DSMLv2 server written in Java.

So we've been consuming open source software for some time, and have been contributing our improvements back to each community. It's been a satisfying process over the years to see our improvements included in new versions of each piece of software.

LWN: Novell has released its UDDI code with a fair amount of fanfare. Can we look forward to other releases of Novell technology in the near future?

Yes, we will definitely release more technology in the future. In fact, we have another open source announcement planned for later in the spring that, like the UDDI server, is related to standards activities. We are also evaluating which proprietary Novell technologies could be good candidates for open source release, although we haven't finalized those decisions yet.

LWN: If I understand correctly, Netware 7.0 will be able to run on top of the Linux kernel. The thinking seems to be that giving customers the option to move to Linux will make them more inclined to stay with Netware. Is that an accurate summation of Novell's strategy? How will Novell respond if it turns out that most customers would rather run on the Linux kernel?

I think it's only one element of our strategy that the option to move to Linux will make our customers more inclined to stay with NetWare. Both versions will be bona fide NetWare 7--whether customers purchase the version that runs on the Linux kernel or the NetWare kernel, they're both revenue-generating products for Novell. If it turns out that most customers would rather run on the Linux kernel, then it would only validate our decision to move NetWare services to Linux. This is the same approach that we've taken with other products, like eDirectory, NetMail, and iFolder.

LWN: Taking it one step further...if Netware 7 runs well on the Linux kernel, what reason would Novell have to continue developing and maintaining its own kernel? What advantage does a proprietary kernel give to Novell when it can run Linux and benefit from the reliability and scalability work being done by IBM, SGI, HP, Red Hat, SuSE, and others?

Novell still has a huge installed base of NetWare customers who depend on a clear upgrade path to the next version of NetWare running on the NetWare kernel. That's why we have a dual-kernel strategy--to ensure that we don't lose customers who want to upgrade to the non-Linux version of NetWare 7. Besides, Linux and the NetWare kernel are both excellent pieces of engineering that have benefitted from years of enhancements and improvements. Many traditional NetWare customers will want the value of the NetWare kernel.

LWN: For customers wanting to run Netware over Linux, will Novell ship a specific distribution, or will customers be expected to obtain a supported distribution from elsewhere?

The answer to this question is in a state of flux. We're not sure yet exactly how this is going to work yet--please bear with us while we sort this out.

LWN: Why is Novell releasing Netware on top of Linux, rather than (or, at least, prior to) Windows?

We're going with Linux because our customers are telling us that they are moving off of Windows and onto Linux. It's as simple as that. Linux has the momentum and the mindshare and we want to lend our considerable energy to Linux.

Comments (4 posted)

Opteron launches

AMD has, at last, released its long-awaited "Opteron" (or "Hammer") processor. LWN does not normally devote much space to following developments in the microprocessor field, but Opteron is worth a mention. There is a good chance that this is the architecture many of us will be running in the future.

Opteron has the potential to deliver the best from both the 32-bit and 64-bit computing worlds. It will run 32-bit x86 code natively, and with good performance. That is a nice feature for people with binary applications, of course, though it is less useful in the free software world. If you have source (and an operating system which has been 64-bit capable for years), support for a new processor is often just a matter of running "make." There is another important aspect to 32-bit support, however: for most applications, 32-bits is the optimal size. Moving to a 64-bit mode involves a sizeable expansion of a program's code and data, with bad effects on cache utilization, virtual memory use, and memory bus bandwidth. Building "cat" as a 64-bit application can only serve to make it bigger and slower. So a processor with native 32-bit support is a good thing.

There are situations, however, where only 64 bits will do. In particular, applications which need to address vast amounts of memory (e.g., big scientific crankers, large databases, emacs) will benefit from 64-bit pointers. So good 64-bit support matters too.

Of course, the thing that really matters for Linux users is Linux support. AMD has worked with the free software community for years to ensure that its processor would be supported. The end result is that you can buy an Opteron server running a stable Linux port (choosing from multiple distributors) today. Windows support, instead, will show up in beta form only later this year, and Apple's support remains a rumor. In some areas, hardware support in Linux still lags behind other systems; with the Opteron, however, Linux got there first. If Opteron lives up to its PR, it could be a platform which brings Linux into many more machine rooms in the next few years.

Comments (16 posted)

Page editor: Jonathan Corbet

Inside this week's LWN.net Weekly Edition

Security: The other security issue; new vulnerabilities in mime-support, snort, ...
Kernel: Memory allocation and interrupt handler changes; internal device number representation; atomic kmaps
Distributions: The State of Regional Linux Distributions
Development: 3D mesh graphics with Gmsh, Apache filters, New versions of OSCAR, CUPS, Etherboot, Glame, Galeon, Mozilla, Vera Fonts, Gmsh, Glade, GStreamer, AbiWord, SQL Ledger, Chandler, KDevelop, Visual Tcl, Treebeard, CUTE.
Press: Will patents pillage open source?, XFree86 Mission Statement, Firebird name collision, SuSE on Opeteron, new OSDL CEO.
Announcements: Linux Household Robot, Open Source Reference Book submissions, EULA vs GPL, Perl Beginner's Portal.
Letters: Desktop Linux; Alternative names for Firebird; Searching for software

Next page: Security>>