Castellers

Third in a series: [1], [2]. I will keep posting results but I can't promise a specific frequency as a fair amount of this (still) involves manual labor.

Introduction

This entry complements other GlassFish adoption indicators reported in TheAquarium and elsewhere, including the Geo Mashup based on GF Admin Console pings (live, geomap@TA), informal Adoption Stories, posts in user lists, and Other Anecdotal Data. All indicators have limitations; my usual comment is that if all the indicators are pointing the same direction, they are probably right.

<Caveat>
• The download numbers quoted below are based on completed (not just attempted) downloads.
• We only count downloads from Sun.Com (under the Sun Java System Application Server name), from Java.Net (GlassFish AppServer) and NetBeans.ORG (NetBeans IDE).
• We are not counting downloads from our Maven Repository.
• We do not count Ubuntu, Solaris, Solaris Express nor OpenSolaris distributions.
• GlassFish is freely redistributable and we don't track other distributions.
• I believe these are solid numbers, but if we discover a problem I will let you know.
</Caveat>

Below are several charts with commentary. The charts show monthly downloads starting in July 2005, right after we launched the GlassFish project, and going through last month, September 2008. The Y axis is normalized to 900K per month to simplify comparisons. The images link to a larger image for further inspection.

Download numbers go down during the (northern hemisphere) summer and at the end of the (gregorian) year. They also tend to peak during JavaOne, in May, and right after new releases.

Corrections and Other Comments

Correction - We corrected severe undercounts in NetBeans downloads since the last report. The scripts we were using made assumptions on the shape of the log files that were no longer valid when we started using a CDN to handle the increased demands.

Known Issues - We believe the new data is correct but, as always, we will report and adjust if we discover any new problems.

Additional Comments - There seems to be a general pickup now that the (northern hemisphere) summer is over. We should also see increased downloads from GlassFish v3 Prelude scheduled for release a the end of this month.

New Charts

As before, there are 3 data sets, collected into 5 charts.

GlassFish AppServer in SDKs

This chart shows downloads of the GlassFish AppServer either by itself (SJS AS 9.x and GlassFish AS) or in bundles like the Java EE SDK and the Java Application Platform SDK (Main Download Page). The peak around May '06 (223,079) was the release of Java EE 5 and GFv1 at J1 '06. Other local peaks correspond to GFv1U1, J1'07, GFv2, GFv2U1 and GFv2U2. D/l numbers for Sep '08 were 220,521.

GlassFish AppServer in Tools

GlassFish is included in, and is the default container for, the most popular bundles of NetBeans 6.1, NB 6.0.1, NB 6.1 and NB 6.5. We only count those from bundles including GF (also see public NB data). NetBeans 6.x has been very popular. The GlassFish portion of these d/l numbers for Apr '08 was 695,550. The corresponding number for Sep' 08 was 507,753.

GlassFish AppServer - Combined

This table just adds the previous two. The combined number for Apr '08 was 882,489 and for Sep '08, 728,274.

JBoss AppServer - From SourceForge

Finally, as a reference point, here is the chart from the Source Forge JBoss AS Download Stats. The d/l for Sep '08 was 112,464. Note - If anybody knows how SourceForge counts downloads (attempted?, completed?), please let me know. Completion ratios vary based on geography and bundle size; for example, GFv2 currently has a completion rate between 60% and 70%. Also I didn't substract the d/l #s (10-15%?) for source bundles or for MD5/SHA-256 files.

Combined Chart

This one just collects the last two charts together. Refer to the previous charts for comments and clarifications, and to the top paragraphs on the value of other adoption indicators.

This is an update to GlassFish Download Statistics from March 6th, 2008. It has an intro - mostly copied from the March entry, some minor corrections to it, and then the new charts.

Added (Sept 25, 2008) - It seems we found the problem with the NB d/l numbers. We are using a CDN to meet the strong demand for NB 6.1 and it seems the script substantially undercounted donwloads in that case. I'll recreate charts when/if the new numbers are available, sorry, no estimated time for now.

Introduction

Readers of TheAquarium will have seen additional adoption indicators like the Geo Mashup based on GF Admin Console pings (live, geomap@TA), informal Adoption Stories, posts in user lists, and Other Anecdotal Data. All indicators have limitations; my usual comment is that if all the indicators are pointing the same direction, they are probably right.

<Caveat>
• The download numbers quoted below are based on completed (not just attempted) downloads.
• We only count downloads from Sun.Com (under the Sun Java System Application Server name), from Java.Net (GlassFish AppServer) and NetBeans.ORG (NetBeans IDE).
• We are not counting downloads from our Maven Repository.
• We do not count Ubuntu, Solaris, Solaris Express nor OpenSolaris distributions.
• GlassFish is freely redistributable and we don't track other distributions.
• I believe these are solid numbers, but if we discover a problem I will let you know.
</Caveat>

Below are several charts with commentary. The charts show monthly downloads starting in July 2005, right after we launched the GlassFish project, and going through last month, August 2008. The Y axis is normalized to 600K per month to simplify comparisons. The images link to a larger image for further inspection.

Download numbers go down during the (northern hemisphere) summer and at the end of the (gregorian) year. They also tend to peak during JavaOne, in May, and right after new releases.

Corrections and Known Issues

Correction - The graphs/data reported last time suffered from some double-counting in Dec07/Jan08 and the charts below show the adjusted numbers. The overall shape of the graphs only changes slightly.

Known Issues - The charts below show a sharp, unexplained, drop in downloads for NB 6.x during Jun 08 and Aug 08. We believe we have found the Aug 08 problem (added some servers but the scripts are not yet searching their logs) but we are still puzzled about the Jun 08 drop (found? - see comment at the top of this entry). I had been waiting for clean data to publish the numbers but I've decided to push it out now; I'll generate new charts when the data for Aug 08 comes, and will let you know if we ever figure out the Jun 08 issue.

New Charts

As before, there are 3 data sets, collected into 5 charts.

GlassFish AppServer in SDKs

This chart shows downloads of the GlassFish AppServer either by itself (SJS AS 9.x and GlassFish AS) or in bundles like the Java EE SDK and the Java Application Platform SDK (Main Download Page). The peak around May '06 (223,079) was the release of Java EE 5 and GFv1 at J1 '06. Other local peaks correspond to GFv1U1, J1'07, GFv2, GFv2U1 and GFv2U2. D/l numbers for Jul08 and Aug08 are 149,435 and 147,575. There is always a d/l lull during the summer, and GFv3 is in a month, so these numbers are as expected.

GlassFish AppServer in Tools

GlassFish is included in, and is the default container for, the most popular bundles of NetBeans 6.1, NB 6.0.1 and NB 6.1. We only count those from bundles including GF (also see public NB data). The GlassFish portion of these d/l numbers for Jul08 and Aug08 are 142,453 and 82,411. We know there is an undercount in August and we should be able to report new numbers soon. I (still) suspect there are also other undercounts but have not been able to prove anything, so, I'm reporting the numbers we have.

GlassFish AppServer - Combined

This table just adds the previous two. The combined number for Jul08 and Aug08 are 291,888 and 229,986. I will also update this chart when/if I get corrected NB data for these months.

JBoss AppServer - From SourceForge

Finally, as a reference point, here is the chart from the Source Forge JBoss AS Download Stats. The d/l numbers for Jul08 and Aug08 are 67,046 and 71,952. Since there might be other bundles not at SF, pointers to additional data would be welcome. Note - If anybody knows how SourceForge counts downloads (attempted?, completed?), please let me know. Completion ratios vary based on geography and bundle size; for example, GFv2 currently has a completion rate between 60% and 70%. Added - it seems that 10-15% of the d/l #s are for source bundles or for MD5/SHA-256 files.

Combined Chart

This one just collects the last two charts together. Refer to the previous charts for comments and clarifications, and to the top paragraphs on the value of other adoption indicators. And we will post updated graphs if / when we can track down the June 08 and August 08 issues.

Continuing with our TheAquarium TV dry runs, the second GlassFish v3 technical meeting is this afternoon at 2pm US Pacific Time (in 2 hours). This is a follow-up to yesterday's meeting, focused on GlassFish v3, specifically, OSGi.

Quoting from Abhijit's email:

Today, we will start at 4 PM Pacific and the topic is OSGi. Sun Santa Clara SCA14/Main Street Conference Room Toll Free: (866) 545-5227 Int'l Access/Caller Paid: 215-446-3648 Access Code: 3535518 The much-rumored and long-awaited wiki with the schedule is here -- http://wiki.glassfish.java.net/Wiki.jsp?page=GFv3EngDiscussion As always, expect changes :-) Thanks, Abhijit

We will broadcast through http://www.ustream.tv/channel/theaquarium; I'll be there are the beginning but Kedhar will be host for most of the meeting.

Join us if you are interested ... there is a GF v3 DEV meeting right now. At http://www.ustream.tv/channel/theaquarium

I have very fond memories of learning to program in Prolog, having learned it from Philippe Roussel while he was visiting Universidad Simon Bolivar. Prolog is a logic language and it was a good mind-strecher after the traditional algorithmic languages.

That's why the series by Hulles showing how to write a Prolog Module for NetBeans 6.1 made me smile (thanks for the tip Charles). Check it out, from Part 0: About the Project to Part 8: Creating an NBM.

So, a toast to Prolog, Philippe, Jorge, all the other USB folks... and to NetBeans! :-)

The experience with uStream.TV during CommunityOne and JavaOne has been very good. Paul Sterk et al. used it during CommunityOne under somewhat challenging circumstances (see Recordings); and the JavaOne keynotes were professionally recorded (Sun GS 1, Sun GS 2, Sun GS 3, Sun GS 4) and look really good. I also did some small tests a couple of weeks before C1 and once this last week and it worked well.

Some years ago I used to host a Sun-internal technology weekly seminar that was quite popular; I had to stop doing it because of other responsibilities but I've been planning to restart it as a global public event as soon as I could find the right technology; uStream seems a good candidate and I want to give it a try. I created a TheAquarium TV channel and I want to start using it in the next few weeks for different events.

Overall, uStream.TV seems to have most wanted features. The setup is very easy for both broadcasters and audience, the quality is quite good and the technology seems to scale well. There are recording / playback capabilities and an online chat with moderation and polling and the host can invite additional hosts.

Below are some of the ideas that we could pursue with TA TV. They are in no particular order; give me your feedback and additional ideas; I'd like to start a few trials soon - I'll announce the trials at TheAquarium and in this blog.

GlassFish Technologies - GlassFish has many important technologies, we probably can do a weekly technical seminar just covering them. The traditional length of these presentations are 1 hour; we could start there and see how it works. Examples of technologes can include:

• Grizzly, OSGi, ....
• Comet, GWT, ...
• Clustering, Failover, ...
• Extensibility, Embeddability, Monitoring, ...

GlassFish Awards Program Series - Presentations based on the GAP Submissions.

MicroTalks or Lighting Talks - This is more a format than a specific topic. We could either have shorter presentations, or take several topics and bundle them into a full hour.

Use Cases and Architectures - Perhaps invite presenters from the Stories program?

Partners - Sun has a new GlassFish Partners Program and we could give them an opportunity to present what they are doing with GlassFish, with some guideance to make the presentations as useful to the audience as possible.

Geo-Focused Talks - One of my goals with this program is to reach outside of my immediate geographic location. That means beyond the San Francisco Bay Area, but also beyond the US, reaching through the globe into all the areas where there is interest for our technologies (see GF GeoMap). Clearly the internet helps but Time-Zones and culture/language may get in the way. A couple of ideas to explore are: having alternate broadcast times (say one for US-Europe, and another for Asia-Pacific), and having events that are hosted by special guests from around the world.

These are just some ideas, please let me know what you think. This is a new medium for me; I'm sure that, with your help, we will learn how to use it as we go through it.

This entry contains some download statistics for the GlassFish AppServer. We usually report this data aggregated once a year around JavaOne but I believe this is the first time we provide some monthly data. Our intent is to automate the reporting and publish it in a monthly basis.

Readers of TheAquarium will have seen additional adoption indicators like the Geo Mashup based on GF Admin Console pings (live, geomap@TA), informal Adoption Stories, posts in user lists, and Other Anecdotal Data. All indicators have limitations; my usual comment is that if all the indicators are pointing the same direction, they are probably right.

<Caveat>
• The download numbers quoted below are based on completed (not just attempted) downloads.
• We only count downloads from Sun.Com (under the Sun Java System Application Server name), from Java.Net (GlassFish AppServer) and NetBeans.ORG (NetBeans IDE).
• We are not counting downloads from our Maven Repository.
• We do not count Ubuntu, Solaris, Solaris Express nor OpenSolaris distributions.
• GlassFish is freely redistributable and we don't track other distributions.
• I believe these are solid numbers, but if we discover a problem I will let you know.
</Caveat>

Below are several charts with commentary. The charts show monthly downloads starting in July 2005, right after we launched the GlassFish project, and going through last month, February 2008. The Y axis is normalized to 600K per month to simplify comparisons. The images link to a larger image for further inspection.

Download numbers go down during the (northern hemisphere) summer and at the end of the (gregorian) year. They also tend to peak during JavaOne, which, nowadays, is in May.

There are 3 data sets, collected into 5 charts.

GlassFish AppServer in SDKs

This chart shows downloads of the GlassFish AppServer either by itself (SJS AS 9.x and GlassFish AS) or in bundles like the Java EE SDK and the Java Application Platform SDK (Main Download Page). The peak around May '06 (223,079) was the release of Java EE 5 and GFv1 at J1 '06. Other local peaks correspond to GFv1U1, J1'07, GFv2 and GFv2U1. February '08 is our second best SDK number so far, with 216,514 SDK downloads. This is very good for a short month outside of the J1 peak season.

GlassFish AppServer in Tools

GlassFish is included in, and is the default container for, the most popular bundles of NetBeans 6.0 and NB 6.0.1 - and also in some older tool releases. NB d/l stats are public; we only count those from bundles including GF. NB 6.0 was released late '07 and NB 6.0.1 last month. They are very successful and their use of GF as their default container has helped the adoption of GF. The February '08 d/l for those bundles is 378,686.

GlassFish AppServer - Combined

This table just adds the previous two. The combined number for February '08 is 595,200. There are two ways to reason on whether or not to add the numbers. On one side, many people using NB do so because of GF; on the other, some are not, either because they use another container, or because they only develop for Java SE or Java ME. At some point we might be able to get more detailed information but, in the meantime, I add the two and keep the original SDK line as a reference.

JBoss AppServer - From SourceForge

Finally, as a reference point, here is the chart from the Source Forge JBoss AS Download Stats. The d/l numbers reported by SF for Feb '08 are 87,215. Note that SF will often return a "Too Busy" the first time you ask for stats; if so, wait a bit and refresh the page. Updated - Note these numbers do not include tool bundles, their enterprise distribution or other JBoss download bundles; I don't have information on the number of additional d/ls from those distributions.

Combined Chart

This one just collects the last two charts together. Refer to the previous charts for comments and clarifications, and to the top paragraphs on the value of other adoption indicators.

I was recently visiting a software group at another company in the area and at some point we ended talking about how information flows in our companies. Their company keeps information flow very tightly controlled; even within the company a group often does not know what other groups are doing. By contrast, I pointed out that I often get my information from Blogs.Sun.Com.

The secrecy or free flow of information is a reflection of the underlying business model: the traditional for-fee Right-to-Use license (which may or not include support) versus the new model based on free Right-to-Use plus optional for-fee support. In the new model the penalty for making most information public is very limited, while the efficiencies of the free flow of information are huge: agility, reusability, connectivity to the customers, architectural review, etc.

There are a few specific cases where it makes business sense to keep some information confidential, like some performance benchmarks, and in some cases privacy may even be mandated, like in pre-released SPEC results, but otherwise, the benefits of public information greatly outweight the disadvantages. Blogs in particular have been extremely useful in addressing large Time-Zone distances and in providing a practical virtual alternative to the water cooler.

Here is another example of the impact of the business model: in the traditional model preannouncing a release might kill substantial revenue but in a subscription-based model there should be no negative impact and actually the opposite may be true! We have often seen sales happen because the future roadmap can help sell the current release.

The move to Open Source has deep implications for the software industry; in more ways that one We are not in Kansas anymore! The next few years will be fun!

I often get asked about what level of adoption GlassFish has, which turns out to be a harder question than it looks and involves some level of tea leaf reading. Below I list some indicators of adoption; none of them is perfect, but they are all pointing the right direction for GlassFish: up!

Downloads - There were at least 3.3M downloads of GlassFish this last year (June-to-June) and the previous year it was over 3M. Traditionally the main vehicle has been the Java EE SDK but now there is also the Java Application Platform SDK and other bundles (SDK Downloads). Direct downloads of GlassFish v2 at Java.Net (GF Downloads) are fewer but growing fast (note to self: create public dashboard). I don't have full information on the competidors, but Apache Tomcat has more downloads (TC Stats@ASF) and, from what I can see, JBoss has fewer (JBoss Downloads@SF). Geo Maps - Downloads don't necessarily mean active users so we started tracking also activity via Geo Maps (want to be the first user in Greenland?). Since the mechanism has several limitations we are in the process of switching to one based on the UpdateCenter, but data is not yet available. In all cases we are careful not to collect private data and we share what we collect.

Surveys - Surveys have several problems; the biggest ones being the sample set and size and the time lag. I track a bit Evans Data; the latest data I'm aware off is pretty old and shows #2 in Linux and #4 Overall. Also see BZ Research from Dec '05 where we were #5, (via RMH and via RedMonk). Hopefully we will see new surveys soon. Another indicator of adoption are Analyst Reports - These have been looking stronger over the last year or so, and even more since we released GFv2 FCS. Check (in reverse chronological order): [Current Analysis], [Gartner], [Forrester] and [Entiva]. Deployment Stories - In a sense, the Growth of the USERS@GF mailing list is a weak representation of adoption/deployments, but we also want more concrete examples and these have been harder to capture than I was expecting. We have a fairly reasonable list at the Stories blog but the large adopters have been hard to pin down for publication (the larger the company, the more people have to approve a public statement). But last week alone I collected 5 leads (and I've heard similar from John, Alexis and others), so I think we will see more additions to that list very soon. Sales, Contracts - Yeah, these would be good indicators but they are also proprietary and/or have customer-confidential data, so, sorry, not for now :-(. Web Searches - So, now we get to the web. There are two obvious metrics. One is number of hits on a search engine like Google; the other is number of queries. Web Page Hits are biased by false positives and by changes in the search algorithms, but one can look at head-to-head searches using things like Google Fights. I did one in March'07 with Friday Fun w/ Google Fights; I'll do an update in a future blog but, because of the limitations, I'm not sure how much of a trend one can capture with this indicator. The other option is to track Keyword Web Searches. This is very easy to do with tools like Google Trends and I had done an earlier version as Fun with Google Trends. This time I'm capturing a few more trends into the graphs posted on the right of this entry but the overall direction has not changed: pretty much everybody's indicator is down, except for GlassFish. From top to bottom: • WebLogic - Live, Snapshot • WebSphere - Live, Snapshot • JBoss - Live, Snapshot • Apache Geronimo - Live, Snapshot • GlassFish - Live, Snapshot • All Together (GTrends only allows 5) - Live, Snapshot • GlassFish and Apache Tomcat - Live, Snapshot Each graph has some biases but the trend is pretty consistent. Also note that there are no units on the vertical axis, so you need to look at the aggregate graph for relative comparisons. Finally, I know that adding "Apache" to Geronimo and Tomcat will undercount but it seems necessary to highlight the trend; just ignore the absolute information and focus on the trend for those. Another nice tidbit from the trend data: look at the Last 30 days, per geo (Live, Snapshot). Right now, it is very nice to see Japan on top; I'd like to think it is related to Restarting the Japanese Translation for The Aquarium Added: I had forgotten the individual graph for JBoss; I added it and resorted the images to adjust for its inclusion.

WebLogic WebSphere JBoss Apache Geronimo GlassFish GlassFish, JBoss, WebLogic, WebSphere, Apache Geronimo GlassFish and Apache Tomcat

We did it! It was a huge effort at the end but the team pulled it off and you can now download both GlassFish v2 and its companion Sun Java System AppServer 9.1. My overview of the release is here; a pretty complete blog rounup is here. The photo shows a Castell being built; like in software, there are Many Roles involved, some are: • pinya - first level support, stabilizes castell, catches anybody falling • folre - a second level support, on top of the pinya • manilles - the third level support, on top of the folre • enxanyeta - the kid at the top • aixecador - the one under the enxanyeta • dossos - the level raising the aixecador • cap de colla - leader, working with advisers and assistants.

The photo shows shows the Margeners de Guissona building a medium-sized castell but here is a big one by the Castellers de Villafranca: a tres de 10 amb folre i manilles (3 at the base, 10 high, all the roles involved). Many people need to do their tasks very well to build such a big castell; like with this release. Congratulations to all!

We released GlassFish v1 in May 2006 and we started working on GlassFish v2 right away. You may have noticed all the GFv2 Release Candidates and the last one didn't produce any show-stoppers so we should be releasing "very soon".

There is a lot of information about GFv2 at TheAquarium and elsewhere; this writeup is to provide an overview of the key points of the release in a single place. Also, since this entry is mostly a bunch of hyperlinks I'm using a glyph () to highlight the links so they are not overlooked.

First, a quick recap from GlassFish v1 (also see here)

• Java EE 5 compliant (Reference Implementation) • Open Source (CDDL and GPL) • Transparent Development (Builds, Mail, Wiki, Bugs, oops!) • Adoption of multiple components by other groups (e.g. Metro and JAXB) • More than 3.3M downloads/yr • Wide Geo Adoption (also see Adoption@TA and Stories) • Included in distros like Ubuntu • Distributed and Supported by Sun as Sun Java System AppServer.

GlassFish v1 was focused on developers, Java EE 5 and on simple single instance deployments; the new release, v2, adds everything needed for the enterprise customer. The main message for GlassFish v2 is:

You don't need to choose between open source and enterprise features,
you can have both

Key New Features in GF v2

Below are some of the main new features in this release; it is a partial list, if you want more information, browse at The Aquarium.

Clustering and HA - GFv2 includes a new in-memory replication mechanism using JXTA for simplified configuration and improved performance, and it inherits HADB from (SJS AS 8.2) for 5-9s availability. See Clustering @ TA.

Commercial-grade Admin/Monitoring - This includes Admin Console, Documentation, CLI Monitoring. All of these have been substantially improved or are new. The Console, for example, has been redone and it is now using the new Woodstock JSF components and JSF-Templating.

All-in One Bundle - The old SJS AS 8.x releases used to come in multiple versions: Platform Edition / Standard Edition / Enterprise Edition. Some of them were free for deploy, others where not. There is now a single, all-in one, moderately sized, bundl (~55MB) that has everything and there is the notion of profiles that is used to adjust functionality based on intent (developer, clustered, enterprise profiles).

Performance - GlassFish is the only open source appserver that has published a SPECj Appserver 2004 benchmark: at the time of posting it was the fastest submission on a single-processor T2000. Performance records are meant to be broken and I believe that there is now a faster 2-processor submission, but we will continue to improve (see related posts @ TA).

Metro WS Framework - GlassFish uses the Metro Web Services framework. This framework includes project Tango (aka WSIT) and the JAX-WS RI and has best-in-class performance and excellent usability. Metro also has very extensive WS-* interoperability with Microsoft which can be used in many applications; a simple example shows an Excel spreadsheet on Vista that has live cells representing content generated dynamically from a GlassFish server.

Tools - GlassFish has very good in NetBeans 5.5.1 support and even better in NetBeans 6 (beta very soon). There is also support for GlassFish in Eclipse 3.3, MyEclipse and IDEA. The GlassFish community also delivers Hudson (continuous integration), Japex (WS/XML testing) and other useful tools.

JBI and Open ESB - GFv2 has built-in support for Open ESB which supports the JBI standard. This includes configuration and administration console.

Modern Web Tier and Scripting - GlassFish v2 supports the latest JCP Web Tier specifications and includes a very complete JSF toolkit: Woodstock. It also supports jMaki to enable consistent encapsulation of common AJAX toolkits and Comet via Grizzly.

GFv2 also supports scripting, including jRuby (Ruby on the JVM) and Phobos (Server-side JavaScript).

Affordable Commercial Support - Sun Java System Application Server 9.1 (aka SJS AS 9.1) is the commercial support brand for GFv2; same code base but controlled patches with bug fixes (description ). We provide support under subscription and purchase at prices that are very competitive and can be purchased directly via the web or through your friendly Sun salesperson. If you have any questions, feel free to contact us at glassfish-business@sun.com.

Adoption - We are already seing a LOT of interest, from the community and the enterprises. The community appreciates the transparency, the quality and the responsiveness of the development team. The driver in the enterprises is quality and support cost; many companies are reassessing their current strategies, and this seems to be accelerating in recent months. Analyst and surveys are lagging but some reports mentioning GF include: Evans Data, Burton Group, Forrester, even Marc Fleury.

Beyond GlassFish v2

GlassFish v2 is one portion of a wider strategy that includes Additional Releases. SailFin moves into the Telco space with Ericsson with higher availability and scalability requirements as well as SIP servlet and other functionality. GlassFish v3 goes across with things like improved jRuby support (like the GlassFish Gem) and down into Tomcat and even smaller spaces like WS gateways and embedded devices.

The GlassFish ecosystem (including partners) already includes community groups, tools, frameworks, solutions, SIs and others. We are already included in a number of distributions and we are pursuing several additional more. Some of this will evolve more fully in the near term, so stay tuned for more info; in the meantime, send us mail to glassfish-business@sun.com if interested.

Last Words

Try it Out! and Stay in Touch! We are always very interested in what works and what does not work. You can stay connected via the mailing lists: USERS@glassfish.dev.java.net and DEV@glassfish.dev.java.net. And feel free to leave comments in this blog entry.

Disclaimers: SPEC and the benchmark name SPECjAppServer 2004 are registered trademarks of the Standard Performance Evaluation Corporation. Sun Fire T2000 (1 chips, 8 cores) 1.4ghz 883.66 SPECjAppServer2004 JOPS@Standard. Competitive benchmark results stated above reflect results published on www.spec.org as of 07/10/06. For the latest SPECjAppServer 2004 benchmark results, visit http://www.spec.org/.

We are maintaining multiple GlassFish release trains to address different user needs, and I put together a summary to help you track them. First some terminology: • Sustaining - In sustaining mode; most changes will be addressing key bugs. • Active - Under active development. • Committed - Committed to happen, schedule may be flexible. • Scheduled - Release with a committed schedule. • Tentative - Release details still not firm.

GlassFish v1 (Sustaining)
• The first final release for this train was J1'06. Other public releases include GFv1 UR1 and GFv1 UR1p1. We are likely to do another public release as soon as we close on GFv2.
• Commercial support is currently available from Sun under the SJS AS 9.0 brand, with releases every 6-8 weeks or so.

GlassFish v2 (Active, Scheduled)
• Adds clustering, profiles, many performance improvements, etc. Published SPECj2004 AppServer numbers.
• Release is planned to go final in a few weeks, in mid-September.
• Commercial support from Sun under the SJS AS 9.1 brand name.

GlassFish v2 UR (Tentative)
• Every major release has an Update Release; it is part of the physics of releases :-)
• Commercial support under the tentative SJS AS 9.1 UR1 brand name.
• Currently pencilled for release at the beginning of 2008.

Sailfin (Active, Committed)
• Adds increased scalability and reliability constraints, as well as SIP Servlet and additional Communication features.
• Tentatively scheduled for mid 2008.

GlassFish v2.next (Active, Tentative)
• GF v2 release based on Sailfin; should include all the improvements in the base AppServer needed to deliver Sailfin.
• Exact packaging relationship to Sailfin still undecided, it is too early to understand the impact of different trade-offs.

GlassFish v3 (Active, Tentative)
• The main feature is a modular architecture based on HK2.
• Services will be added to this train incrementally.
• A key challenge is to maintain usability in the administration while making it modular.
• We may define a web tier/ web services / scripting milestone with limited adminitrative GUI.
• Final release of full AppServer may be aligned with Java EE 6.

Note: All these releases are public open source code. And they will all be supported by Sun via enterprise-grade distributions.

The NY Times Magazine sometimes has very interesting articles. This week it included The Music Man, a piece on Rick Rubin, a very influential music producer that was named co-head of Columbia Records earlier this year. The music industry is in trouble and Rubin's charter is to rejuvenate it. Some excerpts from the article... Reporting on some focus groups with some college students ... a) no one listens to the radio anymore, b) they mostly steal music, but they don't consider it stealing, and c) they get most of their music from iTunes on their iPod. They told us that MySpace is over, it's just not cool anymore; Facebook is still cool, but that it might not last much longer; and the biggest thing in their life is word of mouth On priorites... The most important thing we have to do now is to get the art right...

On the challenges facing the music industry and Columbia...

... we will have the best record company in the industry, but ... we might have the best dinosaur.... until the paradigm shifts, it's going to be a declining business...

Other bits from the article worth mentioning: he started a "word of mouth" department, is a fan of The Beatles (like everybody else?), and I learned about Paul Potts ([YouTube], [WebSite]).

Overall, an article worth reading. Some of the comments about the music industry remind me a bit about the software industry (but there are a number of key differences), and some of the visionary in Rick reminds me of Jonathan :-)

The question may start with "What is the difference between GlassFish and SJS AS?" Or, "What is the difference between how RedHat and GlassFish productizes Open Source?" Or, "Can you provide support for GlassFish?" I've explained this topic a number of times in the last few weeks, so I'm going to write it down and next time I'll just send a URL to this entry... I earlier wrote about Commercial Support for GF, this time I want to explore a bit more the process mechanics.

The first observation is that different customers have different needs and the features of "products" will vary. For now I'll focus on the traditional enterprise customer and on deployment-level support because that is what we are providing today for GlassFish. What such customer wants is insurance that deployments will not break, so what they want is help, early warnings, and bug fixes, but not necessarily the latest features (the same customer may also want the latest features, but that is likely to be from a different group, or perhaps the same people but under a different role).

The way these bug fixes are provided involves creating a source branch under tight control so that only bug fixes are added, and running full test suites, and then delivering a stream of releases addressing the important bugs, with a tradeoff between fixing many bugs and ensuring that there are no regressions. All the companies that I know that provide enterprise-level support for a product do this type of branching, and whether the code base is open source or not does not change that. The lifetime of the code branch depends on how long support is provided: the longer the support the more expensive is to provide it, as the branch and its associated resources (tests, machines, configurations, engineers) have to be maintained active.

A sustaining branch is key to provide support but the main main development will keep moving on. There are different ways to treat these two code repositories, and that varies depending on the relationship between the groups maintaining the two repositories, on the business model involved, and possibly other factors like the length of the releases, etc. A change that occurs in the support repository but does not migrate to the main repository needs to be reapplied the next time a supported release is created, and, if the main repository has changed a lot, this may be very expensive, risky, or just impossible.

In general, the goal is to push the changes from the support repository to the main one but that may not always be possible. In the case of GlassFish / SJS AS, the same groups are involved, so this is not a problem. The bug fixes are pushed through to the main repository where they are combined with new features so later releases will include both features and bug fixes. This arrangement works well for all parties: customers are provided good support and at a good price - otherwise some other provider will use the main repository to provide support - and Sun can get revenue from support.

The last mechanism involved is stabilizing releases to the level expected by enteprise customers before declaring them final. There are two basic approaches about this. In one, the stabilization is done in a separate branch; in the other it is done in the main branch. GlassFish follows the second approach: we stabilize and remove all bugs in the main repository, the final release of GlassFish v2 is going to be identical to SJS AS 9.1.

The GlassFish approach is resource efficient, otherwise we would need to run tests twice and provide separate documentation, etc. Also, for the reasons described above, this approach does not negatively impact our revenue. But this approach implies a substantial level of investment in the project and a buy-in from the community with the enterprise-quality goals for the project. When those are not present, a commercial vendor will need to do the stabilization in a private branch, and then get those changes back into the main branch, with the extra cost and risks associated.

... Does this help? Ask questions in the comments, and, if necessary, I'll provide an improved version later.

I mostly blog at The Aquarium but that format is only useful for short entries. Every now and then I want / need to write longer pieces, so I'm resurrecting this blog for those cases. Now I only need to find the time to do the writing!