Franz Haberhauer's Weblog Franz Haberhauer's Weblog

Nur kurz zwei Hinweise auf englischsprachige Artikel zu Zukunftsperspektiven für Solaris

• Ashlee Vance von The Register , ein eher kritischer Redakteur, sieht die Zukunft von Solaris heute wesentlich positiver als früher
• IDC war vom Projekt Fishworks - einem Framework, das es ermöglicht, einfach Appliances auf der Basis von Solaris zu bauen und im Rahmen der Analysts Conference von Sun Anfang Februar vorgestellt worden war - so angetan, dass dazu ein überaus positives Papier (IDC Opinion) erschien.

Solaris 8 was released in spring 2000. Now after more than six years its end of life has been announced - but this only means that we will stop shipping media kits in February 2007 and that the clock towards the end of service life starts ticking for five years. End of service life will be March 31, 2012. This is a pretty long lifecycle e.g. compared to the seven years offered by RedHat for their Enterprise Linux. With the Solaris lifecycle model Sun has found a good balance between innovation and stability. In the Linux space there is still a discussion about wether such a balance can be found. Kroah-Hartman who works with Chris Wright to maintain the 2.6.x.y kernel patches recently stated "I think the enterprise stable kernel model doesn't work","I think stable enterprise Linux is really an oxymoron". The Solaris model has a much stronger focus on the stability aspects needed for binary distributions whereas in Linux there is a stronger focus on the freedom to innovate in the source code, as can been seen from the deliberate decision to have no stable binary interface for device drivers. Solaris offers the DDI/DKI. For layered software the lifecycles are even shorter. MySQL AB is about to draw the curtain on older databases and they have come up with a lifecycle model offering five years of support.

Linux distribution vendors now recognize that Solaris has taken a lead in one dimension of virtualization - that of an application environment - with its Container technology. Virtualizing the application environment, the application/OS interface is much more lightweight and efficient performance-wise than virtualizing the OS/HW interface. Actually both approaches are complementary and may even be used together to archieve different goals. A CNet News Article summarizes their plans for technologies similiar to Solaris Containers. Yet this will still take some time, whereas Solaris Containers are already deployed in production. For those in the neighbourhood of Stuttgart, Germany the next Solaris@STEP event on Sep. 13 may be interesting at which Jörg Hoss from TDS will talk about their use of Solaris Containers to run SAP environments (in german).

Some conferences, magazines etc. picked up the word Linux as a hyped synonym for "OpenSource". Now they recogize that the Open Source world is actually much wider than just Linux. And Open Solaris get's invited to the party. We have been invited to offer a whole track on OpenSolaris at the german LinuxTag 2006, a conference which despite its name actually adresses OpenSource in general. At this event Andrew Morton, the lead maintainer of the Linux production kernel, gave a talk with got a lot of attention from the press: Linux kernel 'getting buggier,' leader says.

Solaris has some advantages which make a difference - I recently came across an interesting paper Solaris vs. Linux: Framework for the Comparison in Large Enterprise Environments by Dr. Nikolai Bezroukov which elaborates quite a bit more on this topic than most others around.

Jointly with the German Heise Verlag we have put together a DVD for x86 platforms with a special build of Solaris Express Community Release Build 31 for x86, the OpenSolaris sources (build 31), the Sun Studio 11, and Java Studio Enterprise 8 developer tools, a prototype of BrandZ that matches build 31, and a bunch of other contributions including the Solaris Software Companion 1/06. 4GB worth of great software - much more convenient to get than through lengthy downloads. And it includes a great developer environment to be used on (and for) the OS it comes with or on a plain Solaris 10.It comes with the 4/2006 issue of iX-Magazin, a leading German UNIX-oriented publication. To follow up on this my team, Sun's german OS-Ambassadors, has set up a group blog (in german): Solarium. P.S.: Kudos to Detlef Drewanz - he actually did most of the work :-)

Jointly with AS-Systeme, our local Sun Education Center, I'm running Solaris@STEP, a series of free workshops on Solaris (in German!). We already covered DTrace and ZFS.

Schedule:

• 14.12.2005 Solaris 10 Container
• 09.02.2006 Solaris Performance

For details and registration see Solaris@STEP

Today we started the Sun SPEED64 Roadshow 2005 roadshow in Germany featuring our new AMD Opteron based servers Sun Fire X2100, X4100, and X4200 and our software offerings for these platforms.

I am giving a presentation on Solaris 10 for developers which will be posted by marketing soon. For immediate access here a list of references from my presentation [Update: Slides available now, the list below includes some additions] :

DTrace

• Statically Defined Tracing Probes
• AnswerBook Chapter "Statically Defined Tracing for User Applications"
• DTrace and Java
• DTrace Tool Collections: /usr/demo/dtrace, OpenSolaris DTrace Community, DTrace Tools Collection by Brendan Gregg
• Enabling User-Controlled Collection of Application Crash Data With DTrace

Security

• Solaris Security for Developers Guide
• Signed Binaries
• Solaris Secure Execution

Performance

• Multithreading in the Solaris Operating Environment

Availability

• Identifying Memory Management Bugs Within Applications Using the libumem Library
• Adam Leventhal on libumem
• Hidden features of libumem - firewalls and protecting overflows by Peter Harvey

Update: More Useful URLs

• Bringing Your Application Into the Zone
• Solaris Service Management Facility - Service Developer Introduction
• Liane Praza: SMF - Administration and Development
• SMF Documentation

Overview of new Solaris 10 features with many references.

A discussion with a customer reminded me of a discussion thread in the OpenSolaris DTrace forum on "a bit of IBM anti-DTrace propaganda" which especially points out DTrace' superior security model compared to the approach taken in SystemTap - which is still work in progress for Linux. The IBM paper positions a whole suite of (future) features in Linux on Power versus DTrace ignoring the big advantage you have from a single tool when you need to correlate data from various sources. They actually ignore features available on the Solaris platform beyond DTrace, e.g. the "Post-Link Optimization" possible with Sun Studio 10 - this feature (Profiling and link optimization (B2.100)) has been available on the SPARC platform since many years when previous versions still had been called Workshop.

The Robert Frances Group released a study on "TCO for Application Servers: Comparing Linux with Windows and Solaris" (right now this URL does no longer work) generating headlines stating Linux 40% Less Expensive Than Windows, 54% Less Expensive Than Solaris. It has been sponsored by IBM and e.g. Tom Sullivan already analyzed some of this report's flaws as he warns to never take vendor-sponsored reports at face-value.

When I read this headline I wondered why Solaris should be more expensive than Linux, given that Solaris service and support is priced lower than either Windows or Linux - a fact Tom Sullivan found with surprise even in this report. The reason is pretty simple: The report included hardware cost and based their calculation on list prices. For Solaris they only considered the SPARC platform, though it would have been easy to compare the various operating systems on exactly the same x86-hardware, eliminating assumptions made regarding actual pricing. This would have made Solaris really compelling. In addition keep in mind that there is no price for Linux per se, there can be a significant difference wether you pick a SuSE or RedHat distribution. When comparing subscription prices you really need to make sure that equal service levels are compared, as e.g. Novell/SuSE charge extra for phone support.

Another aspect of this study may also be considered a systemic flaw: the focus just on the application server environment. Typically it is part of a larger environment. It makes a difference in TCO wether you run such a multi-tier environment homogenously or heterogenously - backend systems are the domain of RISC-based servers and a homogenous Sun SPARC solution may have a lower TCO than a heterogenous solution combining AIX/POWER and RedHat/x86. Related to this is the cost for administration staff. While such TCO studies often correlate it with the operating system actually it may be more correlated to the complexity of an environment.

On the hardware side there may be major differences among different x86-Plattforms. Power and cooling have become major cost factors these days. Marc Hamilton did an interesting analysis for a large compute cluster comparing Intel- and AMD-based servers.

Over the next months Sun will release products which are going to disrupt existing TCO studies.

Here you may find a very nice collection of DTrace tools and examples.

To fully exploit the capabilities of DTrace some understanding of Solaris internals is useful. Jim Mauro and Richard McDougall have put a current (March 2005) 536 slides presentation on Solaris internals and performance on their website www.solarisinternals.com which complements their book. While the book is a little dated (reflects Solaris 7) it's still a good basis and you may use it together with the slides mentioned above while they are working on the second edition. In addition their blogs offer a wealth of articles on how to use DTrace to understand what's happening in the kernel.

Alan Hargreaves has put together a nice article on how to use Statically Defined Tracing (SDT) Probes to instrument your own code such that you can use the DTrace framework for monitoring and analysis. SDT is documented in the DTrace manual, but it's a rather thick manual and it's likely that many software developers will miss this promising use case: a powerful instrumentation for your own code tightly integrated with the instrumentation in the OS - something performance analysts will highly appreciate - without generating overhead if not active.

Casper Dik gives a glimpse on what's ahead in Solaris on x86: The new boot system, ACPICA (Advanced Configuration & Power Interface Component Architecture) and wireless support. It will still take a few months until all of this will show up in Solaris Express and Solaris 10 updates, but we are on the way to make Solaris a decent OS for your laptop - for many purposes it already is. Alan Zeichick confirms in a Solaris 10 review on Infoworld that Solaris 10 has made real progress in device support on the x86 platform.

Yesterday we announced the first million of registered licenses for the free Solaris 10 OS. Since January 31, 2005 Solaris 10 can be downloaded and since early March media kits are available for customers with support contracts or for purchase.

At CeBIT, with 680.000 visitors and nearly 6.300 exhibitors world's largest IT fair, Solaris was one of the highlights at the Sun booth. We distributed more than 3.000 kits with Solaris DVDs (both for SPARC and x86). A few of these are now being auctioned at eBay.de (e.g. this or that) - before bidding keep in mind that unlike the full media kits they do not include e.g. the Answerbook whose content may be found on docs.sun.com or the Software Companion CD, which may be downloaded from here.

Though the official release name for the current version is Solaris 10 3/05 (based on the availability date of the media kits), there are even more recent bits available of the current development version - codenamed "Nevada" - through the Software Express for Solaris program which has been highly successful during the development of Solaris 10 and which is beeing continued while we move towards Solaris 10 updates. This week the second of the monthly releases is due. Here some information on its contents. So the Solaris Express 2/05 and 3/05 versions are actually more recent than the commercial Solaris 10 3/05. Confused? Here some more background how these Solaris offerings are related.

Sun has made the next step with OpenSolaris. We have announced the licensing, we have released the first piece of code: the source code for DTrace, one of the key innovations in Solaris 10, and we have disclosed the roadmap for OpenSolaris. Users in a pilot program do already have access to the full buildable source and actually have already built OpenSolaris from it, e.g. Pieter Van den Abeele or Jörg Schilling. Ben Rockwood, another member of the OpenSolaris pilot program and the supposedly the first who built OpenSolaris externally, maintains an extensive blog. A whole list of bloggers on OpenSolaris may be found here.

We went with the Common Development and Distribution License (CDDL), an OSI approved open source license, which evolved from the Mozilla Public License (MPL). In 1999 David Wallace Croft compared the GPL and the MPL and elaborated why from his perspectives the MPL is actually the more free license as it does not have the "viral" nature of the GPL and allows more business models.

Update: For a current discussion of aspects of the CDDL versus GPL see Pamela Jones' comments at Groklaw and Simon Phipps' response

The first piece of code which is featured on www.opensolaris.org are the sources for DTrace. Bryan Cantrill and Adam Leventhal from the DTrace development team provide quite a bit of background information in their blogs.

Within the first 24 hours more than 50.000 unique visitors viewed www.opensolaris.org and there have been more than 1.500 downloads of the DTrace sources in the first 12 hours. Pretty impressive. Jim Grisanzio, Sun's Community Manager OpenSolaris, has collected a lot of feedback on the OpenSolaris launch.

Martin Fink, HP's Linux Vice Precident, started blogging with an article on Solaris - this at least shows the significance of Solaris. In this article he tries to spread fear, uncertainty, and doubt (FUD) based on poorly researched statements and assumptions on Solaris and Sun's strategies. Ben Rockwood issued a response from the Solaris users and developers community and Alan Hargreaves, an australian Sun engineers, also put together a response. Simon Phipps did some forensics of Martin Fink's web site with interesting results.