trip report DECUS europe symposium 1997
during the symposium, this trip
report was updated on a daily basis.
(last update: october 20, 1997, 10:00, final version)
my family took me to the airport zürich on saturday, october 4, 1997. actually not the whole family, my oldest son daniel said it is too sad to say good bye at the airport and he decided to stay at my wife's sister's house.
when i waited for the bus to bring us to the plane, i heard the first friendly "hello". it was "gigi" who was also on his way to the DECUS europe symposium.
the plane took off on time towards rümlang, the village where i
grew up. i saw the campus of the ETH hönggerberg where a giant new
building is under construction. then we turned around and crossed the airport.
we flew over winterthur, frauenfeld and konstanz. the weather was beautiful, we
could see the alps in the distance. the valleys were filled with a light fog.
what a wonderful country from above !
soon after stuttgart, we saw the first clouds and minutes later, we couldn't hardly see any ground at all. in copenhagen, the whether was cloudy and windy. our jumbolino moved like a leaf in the wind, but the pilot managed to bring us down so smooth that i couldn't tell when the plane touched the ground.
when we left the plane, i told "gigi" about one of our suite cases that once was taken by mistake by another passenger in L.A. while we both were laughing, we realized that "gigi's" suite case didn't made it to copenhagen ... after having filled in the required papers, we run into neil gregory and together we took a ride into the city of copenhagen.
at 19:00, i had an appointment with pierre-yves baumann at his hotel. unfortunately, i forgot to write down the name of his hotel - and there was no Internet access to read the mail. i was positive he stayed at the "grand hotel" and i waited there until 19:20, but he didn't show up. therefore i went back to the old downtown, where i run into him just by accident half an hour later ! he didn't stay at the "grand hotel" of course, but we had a great diner anyway.
06-oct-1997: since the symposium just has started,
there is not too much to say yet. the registration worked out fine for all of
us as far as i know. however, most attendees dislike the fact that the
conference center bella center is far away from the center of
copenhagen. it takes about twenty minutes by
bus from the main station to bella center. depending on how far the hotel is
away from the main station, it may take quite a while to get there.
even though the symposium fee includes a ticket for local public transportation, this ticket is NOT included in the material one receives at the registration desk. instead, one has to ask for the ticket in a separate office. i consider this a very strange and unfair policy of the local organization committee !
there are about 50 PCs available at the Cyber Cafe - running Windows NT of course. the Internet connection on sunday was striking fast, i hope it will be the same for the rest of the week ...
07-oct-1997: well, the network is still doing ok - except for the fact that some idi** has decided to disallow ftp all of a sudden. and today, the local public transportation system has proven to be a disaster. the DSB seems to have no idea what it means if hundreds of people wish to attend a symposium and are forced to use public transportation to get there. when i waited with a lot of other attendees in front of the main station, the bus did not only arrive late, but it was also so crowded that it did not even stop there. we waited close to half an hour until we could finally get on a bus even though according to the schedule there should be a bus every ten minutes. this is not only a shame for the local public transportation organization, but it also proves that it is always a bad idea to choose a location where the attendees rely on public transportation to get to the symposium and back to the hotel. i certainly hope that DECUS europe in the future will choose locations where the conference center and the hotels are within walking distance !
09-oct-1997: today was the last day of the symposium. unfortunaly, there was no formal closing session, so most attendees started to disappear during the afternoon. i'm now sitting here in the hotel in copenhagen and ask myself:
well, the first question is easy to answer (i acctually gave the answer on monday already): YES ! the training seminar about the internal architecture of Windows NT was worth the trip just by itself. but there was other interesting information as well. so from my point of view, the main messages are:
paula sharick from wildwood associates conducted a guided tour through the Windows NT registry and described tools and methods to look at, modify, backup and restore the most important system database of that operating system.
the registry is a replacement for the various INI-files that were so
popular under Windows V3.x. it can also be seen as an extended version of
OpenVMS' SYSGEN parameter file. it is a database that describes the whole
system, its hardware and software components and all user profiles. it may
contain hundreds or even thousands of parameter name and value pairs, organized
in five major trees or hives, as they are called.
but only HKEY_LOCAL_MACHINE and HKEY_USERS are real hives, all others are just different views of the same data. hives are backed up by a number of files. to duplicate some characteristics of one Windows NT system to another (e.g. shares), it is possible to save the information on the source system into a text file by using the registry editor, copy this file to the target system and apply the information to the registry of that system.
two interesting programs paula mentioned are:
probably the most important part of paula's talk was about backing up the registry files. the main problem with those files is the fact, that some backup tools - such as NTBACKUP - do NOT copy open files. but some files of the registry are always open and therefore won't be backed up by such a program. paula suggested either to use third party backup software or tools that come with the SDK (Software Development Kit), such as REGBACK and REGREST.
IMPORTANT NOTE: manipulating the registry may destroy crucial information and may leave your system in an unusable state !
david solomon presented this seminar, which was attended by about 50 persons. this seminar alone was worth the trip to copenhagen and all the money the ETH is going to spend for this symposium ! he gave a very very interesting talk about the internals of Windows NT. it is by far too much information to be presented here in detail (the handouts contain 316 slides !), but from my point of view, these are the highlights:
david pointed out many tools that are either on the workstation or server distribution, on the SDK or DDK. in addition, these are some of the very useful websites related to Windows NT:
i'm very sorry that i missed the opening session due to the extremely bad performance of the local public transportation DSB !
i also missed part of the first keynote session due to the extremely bad performance of the local public transportation DSB !
however, i heard at least part of hans dirkmann's talk. he said that DIGITAL is doing well again, but that his company has a problem with the reputation of its brand name. they will work on that to improve it and he mentioned a few success stories which should prove that DIGITAL is on the right path.
samuel fuller stressed the point
of the year 2000 problem. he said, it is a
serious, big, but manageable
problem. DIGITAL's hardware is already y2r (year 2000 ready). OpenVMS and UNIX
will be y2r by the end of this year. DIGITAL will even provide fixes for
OpenVMS V6.x and V5.5-2 to make them y2r.
of the more than 1500 layered products, 711 are already y2r, 345 are planned to become y2r, 470 are still under evaluation and 125 products will be retired before the year 2000. DIGITAL's goal is to have made all decisions whether they are going to make a product y2r or not by the end of 1997.
DIGITAL has set up a number of competence centers to help their customers to make their applications y2r.
samuel fuller said, the year 2000 problem is a good example of the major challenges that the computer industry will face in the near future, such as the introduction of the euro or IPv6.
finally jesse lipcon let the 20 years of the history of VMS pass by. the main message is
jesse lipcon started his journey in 1977 with the announcement of the VAX 11/780 and VAX/VMS V1.0. the first VAX was actually delivered in 1978 to the carny milligan university. in 1983, the first VAX on a chip was announced as the microVAX I. jesse lipcon mentioned two numbers that i found very impressive:
very briefly, these are the highlights of the Windows NT update and future sessions:
in the more distant future, there will be some cluster functionality in Windows NT. but these capabilities will only provide failover functionality and will not be as extensive as with OpenVMS.
the most important message is of course: it is alive and it is not going to die !
there are two major issues in the future of OpenVMS:
OpenVMS V7.2 is to be expected between middle and end of 1998. there will also be a V7.1-H1 to support additional hardware such as improved SCSI features and memory channel 2 (a fiber based memory channel). in the more distant future, there will be support for ATM and Oracle V8.
DIGITAL UNIX V4.0D, which will be shipped in december 1997, will provide the following enhancements:
DIGITAL UNIX V5.0, which will be shipped between middle and end of 1998, will provide the following enhancements:
during this talk, someone mentioned a very interesting software utility called FreePort Express. it deploys the same technics as FX!32 but translates binaries compiled for SUN systems into Alpha native binaries that can be executed under DIGITAL UNIX. this utility translates also programs for DIGITAL Ultrix compiled for MIPS systems, other binaries may also be translated ...
the social event took place in the Øsknehallen,
probably an old farmer's market. there was a drawing of a brand new DIGITAL
HiNote 2000. the winner was a mr. meier, most likely from germany. the
gentleman grabbed the notebook, posed for a picture and disappeared without
saying a word. after this, some members of the DECUS europe board celebrated
the 20th anniversary of VMS by cutting the birthday cake. both, the drawing and
the celebration could have been performed a little bit more spectacular, but
they obviously decided not to do so. the food was good, however, the policy of
the payment for the drinks was again quite strange - as with the bus tickets.
if one ordered a drink at the bar, one had to pay, if one ordered a drink at
the table, it was free ... but only up to 10 pm ...
the performance of the Café Kølbert (ten waiters and their mistress) was quite entertaining. the act of the Tivoli Pantomime Theatre was probably appreciated by the lovers of ballets. i personally was more delighted by the performance of the New Orleans Delight. on the way back to the hotel, copenhagen greeted us with a little rain shower ...
the three main concerns of the affinity program are:
as part of the affinity program, the following products or enhancements will become available:
ken coar explained two mechanisms to restrict access to web pages:
if i understood correctly, the following representatives of DIGITAL tried to answer the questions from the audience (actually, the questions had to be filled into a form so they could be sorted out, but the attendees still had the oportunity to ask questions live):
Q: there were articles in the newspaper this week,
saying that DIGITAL is going to sell its Alpha processors business. is this
A: DIGITAL is committed to the Alpha product line in the short and in the long term ...
Q: how reliable is the DIGITAL - Microsoft
A: jim suggested to involve Microsoft in the DECUS activities to ensure a good relationship.
Q: everybody talks about Windows NT these days. has
DIGITAL abandoned the installed OpenVMS and UNIX
A: jessi said: "we have three children and we love all of them". around 2500 people work in the OpenVMS and UNIX engineering, this should prove that DIGITAL is still committed to these operating systems.
Q: the base products are there for Windows NT on Alpha,
but the nice tools are only available for the Intel platform. what will DIGITAL
do about this ?
A: with Windows NT V5.0, a new component called WX86 will be included and fully supported by Microsoft. this allows the execution of Intel code on Alphas with reasonable performance. this software will be included in the beta 2 release of Windows NT V5.0, which should become available in january 1998.
Q: do the Microsoft engineeres have no Alphas available
or what is the reason that they seem not to care about the
A: "Microsoft is a business driven company" (note from the author: i have heard the same phrase before somewhere ...). DIGITAL will make sure that the important products will become available as native Alpha code. Visual Basic is the next product, that will become available on Alphas, others - such as the Java studio - will follow.
Q: is DIGITAL going to deliver boxes only in the future
and stop developing software ?
A: DIGITAL will continue to develop software, but they will focus on web bases services, clustering, software development products etc.
Q: are AlphaStations running OpenVMS still a good
choice in the future ?
A: the focus for OpenVMS will be on the server side in the future, whereas desktop system most likely will run Windows NT ...
the swiss attendees gathered at 18:00 at the both of HIS Technologies AG, the sponsor of the apero. i had the great pleasure to meet an old friend named paul shrager. unfortunately, paul santner could not join us due to important business issues, but at least mr. ralf brandmeier was there. also franklin dallman and his wife honored us with their presents.
at the restaurant nyhavns færgekro we had some sort of diner (i don't like neither seafood nor fish, but the desert tasted great :-). we were a nice party of 36 people and despite the strange kind of food, i enjoyed that event a lot. i even heard that there will be a stammtisch in the eastern part of switzerland in the very near future, so please watch our homepage for details. and since DIGITAL switzerland payed for the food and beverages, i'm now a little bit more convinced that it is true what hans dirkmann said yesterday: this company is on the right track again. thank you sonja koch for all the work and thank you DIGITAL for sharing this evening with us !
DIGITAL showed quite an impressive amount of hardware. i liked the new DIGITAL HiNote 2000 with its 14.1 inch screen a lot . the various StorageWorks boxes provide RAID capabilities and other features to improve high reliability and availability.
there were two very interesting talks about TCP/IP V6 by yanick pouffary and robert watson. the later is a colleague from the old days when i used to be a DEC employee ...
yanick explained the motivation for IPv6 and its main features. address shortage is not the only reason for a next generation of TCP/IP. other issues are performance and security.
bob talked about the practical aspects of IPv6. there is a IPv6 based network in place. many vendors have working code available for IPv6. those implementations of the TCP/IP stack support both versions, V4 and V6. the transition phase is expected to take place over many years. IPv6 is designed to be able to coexist with IPv4 implementations. bob showed a couple of cases, where it is possible and reasonable to deploy some of the IPv6 capabilities today. he expects, that especially in europe, there will be a great demand for new IP-addresses with the introduction of the de-regulation of the telecommunication. new ISPs (Internet Service Provider) will pop up everywhere. these companies might be very interested in IPv6, because it supports for example roaming much better than IPv4.
for further information about IPv6 and 6bone (the IPv6 based Internet), checkout the web.
from my point of view, this was certainly the most interesting session today. the OpenVMS registry will first become available with the field test of OpenVMS V7.2. the major role of the registry is to act as a configuration database for applications which are based on DCOM (Distributed Computing Object Model). there will be a single clusterwide registry with failover capabilities. the registry will be backed up by files such as REGISTRY$LOCAL_MACHINE.REG and REGISTRY$LOCAL_USER.REG which will be stored in SYS$NTA (NT Affinity). modifications to the registry will be transaction based, so there will be some additional LOG files. access to the registry will be provided locally through a command line interface and remotely from Windows NT through the well known registry-editors. the later will require PATHWORKS V6.0 or later on the OpenVMS system.
the registry will neither replace SYSGEN nor the IO database. it is therefore NOT required to boot an OpenVMS system. nevertheless, the main goal of the implementation of the registry on OpenVMS was to make it bullet proof. with other words: it shall not be possible to corrupt this database.
the registry can either be accessed through a new system service called SYS$REGISTRY[W] or through 3rd party WIN32 APIs that have been ported to OpenVMS. the registry requires UNICODE support, so this has been added to OpenVMS V7.2 as well.
one major open issue with the registry implementation is how should the cluster structure be reflected in HKEY_LOCAL_MACHINE.
from the audience, there were suggestions to add lexical functions to query the registry and to add access to other important system information - such as the SYSGEN parameters - through the registry.
i have hardcopies of the documentation from the two training seminars i have attended. all the slides of the sessions are available on a CD-ROM, additional information and updates may be found on the DECUS Europe web server.
trip report DECUS symposium 1996 in barcelona, spain
|to the ETH homepage||to the DECUS homepage|