Conference Program Thursday February 15
Speaker: Stephane Eranian, Hewlett-Packard Laboratories Palo Alto CA
The IA-64 architecture, co-developed by HP and Intel, and its first imple-mentation, the Itanium (tm) processor is going to reach market sometime in 2001. For the last two years HP Labs and other major companies have been working to port the Linux operating system to this new 64-bit architecture. The full source code produced by the project was contributed back to the Open Source Community in February 2000 when most of the specifications became public.
In this presentation, we first describe the key architectural features of IA-64. Then, we give a short history of the project: how it started, who is involved and where it is now.
The next section covers the kernel internals. We describe some of the key subsystems including virtual memory, system calls, process state and IA-32 emulation.
Finally, we give an update on the user level status for compilers, C library, math libraries, X11, Java as well as standard Linux distributions. We also describe the IA-64 Linux Developer's Kit which allows Linux/IA64 developments on a regular Linux/x86.
Speakers: Brian Pawlowski, Network Appliance Inc. and Spencer Shepler, Sun Microsystems
Brian Pawlowski is Chief Technical Officer at Network Appliance Inc., where he works on scalable storage architectures. He is co-chair of the IETF work-ing group tasked with the design of the next version of NFS. He has recently been studying the problem of utilizing Gigabit Ethernet bandwidth with distributed file systems. Before joining Network Appliance, Brian was at Sun Microsystems, Inc. where he worked on distributed file systems. He was one of the principal engineers involved with the definition of NFS Version 3. He also worked on file server performance in symmetrical multiprocessing architectures.
Spencer Shepler, who has worked for Sun Microsystems since 1998, is the edi-tor and co-author of the NFS version 4 protocol. Spencer has been in the NFS industry for 11 years. First with IBM and then with Sun. Spencer has also been involved with the development and release of both versions of the SPEC NFS benchmark. Spencer’s current responsibilities are leading Sun’s NFS version 4 development team along with involvement in the IETF NFS version 4 Working Group.
The Network File System (NFS) Version 4, is a new distributed file system similar to previous versions of NFS in its straightforward design, simplified error recovery, and independence of transport protocols and operating systems for file access in a heterogeneous network. Unlike earlier versions of NFS, the new protocol integrates file locking, strong security, operation coalescing, and delegation capabilities to enhance client performance for narrow data sharing applications on high-bandwidth networks. Locking and delegation make NFS stateful, but simplicity of design is retained through well-defined recovery semantics in the face of client and server failures and network par-titions. This talk will reveal all.
Speaker: Rik van Riel, Conectiva S.A.
Rik van Riel first installed Linux in 1999 (with lots of help from friends). Despite his late start, with Slackware 2.x and kernel 1.1.59, he has mana-ged to catch up enough to be an active contributor to various Linux projects.
Besides being a kernel hacker, he helped along a few documentation projects, he co-founded NL.linux.org and is one of the hosts of the #kernelnewbies IRC channel on irc.openprojects.net.
Thanks to Conectiva S.A., a big Linux company from Latin America, he is now able to practice his hobby full time. Currently his main activities are hacking the VM subsystem, maintaining the fair scheduler and assisting other kernel hackers on #kernelnewbies.
16 This lecture will start with a (short) introduction to memory management. A quick overview of the problems that virtual memory management faces, and some tactics to cope with them will be presented.
After that we will take a look at the past and present of the memory manage-ment subsystem in the Linux kernel. We will see how Linux VM works and why it does not work very well in some cases.
The lecture will conclude by analyzing the strong and weak points of both the thing presented in the VM introduction and older Linux VM code. This leads to the design for the new VM, which is currently being written by Rik and others.
Speakers: Jamal Hadi Salim, Nortel Networks and Robert Olsson, SLU
Jamal Hadi Salim has been working at Nortel Networks for over six years in a variety of areas ranging from Digital Signal Processing to SS7. He has been contributing to Linux since 1994.
Robert Olsson is deeply involved in the Bifrost project, which distribute a small Linux flashdisk-distribution for routing/firewalling/IP-login, which has a lot of user, but are still quite anonymous.
The availability of cheap, powerful commodity hardware makes it more viable to build routers based on Linux. Given the Open source nature of Linux, One can deploy a lot more features on Linux in a shorter time than on a commercially based one. In fact, Linux has already got a lot more features than a lot of expensive commercial routers.
We attempt to enhance Linux’ capacity for routing. Initial attempts indica-te a doubling of routing performance over vanilla Linux on a single pro-cessor x86 PC. In this presentation we reveal some of the techniques used to scale Linux.
Commodity hardware with a well-tuned operating system can today deal with current Internet loads in many installations and it provides an extremely flexible platform for routing.
Routing and network capabilities of commodity hardware combined with new software techniques is now becoming an emerging area of research.
Speaker: Poul-Henning Kamp
Poul-Henning Kamp believes that UNIX is the best OS ever made so far, he is convinced we can still make it better and he has been trying to do so since the early eighties.
Ever since Minix 1.0 came out, Poul-Henning has been running UNIX on his laptop, and via 386BSD he came to FreeBSD. In FreeBSD he has been release engineer for a number of FreeBSD releases, written, rewritten and cleaned up many pieces of FreeBSD kernel, written a memory allocator, a password scrambler, the beerware license and generally been having a good time doing so.
Poul-Henning lives in Denmark with his son, his daughter, about ten FreeBSD computers and a couple of the world’s most precise NTP clocks. He makes a living as an independent contractor doing all sorts of magic with compu-ters and networks.
Which operating system is the power behind major Internet sites like The Internet Movie Database and Yahoo? Which OS is inside the big routers from Juniper and Ericsson? Which OS is used in embedded products by the likes of Nokia and IBM? What does Slashdot put up as a firewall when they are under attack? The answer to all these questions is “FreeBSD”. FreeBSD is closing in on release 5.0, which will contain significant new features and improve-ments.
This talk will present the latest news from the FreeBSD project.
Speaker: Jes Sørensen, Linuxcare Inc.
Jes Sørensen has been working on the Linux kernel for more than six years, the last three as the maintainer of Linux/m68k. He used to work at the European Laboratory for Particle Physics (www.cern.ch/) where he worked on very high performance networking, Linux clusters and Linux/IA64. This has included writing Linux device drivers for Gigabit Ethernet and HIPPI (High Performance Parallel Interface, an 800Mbit/sec supercomputer network). Jes now works for Linuxcare Inc. in Canada (www.linuxcare.com) where he continues to work on Linux/IA64, high-speed networking and other low level Linux issues.
Linux is moving into the world of supercomputing and has started to take over jobs previously reserved for supercomputers and high end workstations from companies such as SGI, CRAY, DEC and IBM. On top of this, gigabit class network equipment, in particular Gigabit Ethernet, has become a commodity item. This makes it feasible to integrate Linux systems into existing super com-puter network facilities and create new Linux based gigabit class networks such as in clusters and for servers. Recent developments such as the Tux web content accelerator, which holds the current specweb99 record shows that Linux is to be taken serious in the world of future networking.
This talk will discuss the history of Linux high speed networking and the author’s work on Linux drivers for HIPPI (HIgh Performance Parallel Interface) and Gigabit Ethernet as well as general issues on how to optimize and tune applications to benefit from high performance networks. It will continue with a look at the next generation of high performance network improve-ments features in Linux such as zero copy networking.
Speaker: Frederico Mena-Quintero, Helix Code Inc.
Frederico Mena-Quintero is one of the co-founders of the GNOME project. He is employed by Helix Code Inc., where he is the coordinator of the calendar components in the Evolution groupware suite. Federico is also the maintainer of several other packages and documents in GNOME and was the maintainer of the GNU Image Manipulation Program (GIMP) before he started with GNOME.
One of the goals of the GNOME project is to address the lack of certain modern technologies in free systems. We have created a number of libraries that make it easier to write large-scale applications. This presentation will show how the Evolution groupware suite uses the following features in GNOME: the Bonobo component model, the GNOME canvas display engine, the gnome-print libraries, the GtkHTML component, the Camel mail-handling library, and how all of this is integrated together to create a good experience for the user.
Page maintaned by Jan