Tuesday Dec 23, 2008

Spiffing up the GlassFish Admin Console

Jason has posted a couple of blog entries (here and here) on spiffing up the GlassFish admin console. We do get a \*LOT\* of positive feedback on the GlassFish administration console, both in feature-richness and ease-of-use. Jason and team are off to make a good thing even better, so feel free to provide feedback or ideas of your own to the team.

You'll see on these threads that I'm a bit of a party pooper. Jason and team have delivered a great console that a lot of folks like, and I worry that making too much of an improvement may interrupt those who feel comfortable with the existing console.  Please provide feedback to Jason and team. The more the merrier.

Monday Dec 22, 2008

GlassFish Enterprise Server: Outstanding Price/Performance

After taking a look at Tom Daly's post on compelling price/performance, I thought I would take a slightly different approach. Tom's post covers the details of the recent 100% open source software result of 1197.1 JOPS@Standard using GlassFish, OpenSolaris and MySQL running on a Sun X4150 Server. The competitive results are also on quad-core X64 servers.

The data shown here includes 3 years of 24x7 phone support for each operating system, database, and application server.  This provides a more appropriate scenario for those deploying in a production environment. While Tom's post has the raw data, I'm just showing the updated $/JOP.

 Application Server

GlassFish Enterprise Server $33.90/JOP
Oracle WebLogic Server Standard Edition 10.3 $232.68/JOP
Oracle Application Server 10g Release $193.71/JOP

Required disclosure : SPEC and SPECjAppServer are registered trademarks of Standard Performance Evaluation Corporation. Results from www.spec.org as of 11/05/2008. 2xSun Fire X4150 (8 cores, 2chips) and 1xSun Fire X4150 (4 cores, 1 chip) 1197.10 SPECjAppServer2004 JOPS@Standard; Best result with 8 cores in application-tier of the benchmark: 1xHP BL460c (8 cores,2chips) and 1xHP BL480c (8 cores,2 chips) 2056.27 SPECjAppServer2004 JOPS@Standard; Best result with 2 systems in application-tier of the benchmark: 2xDell PowerEdge 2950 (8 cores, 2 chips) and 1xDell PowerEdge R900 (4 chips, 24 cores) 4,794.33 SPECjAppServer2004 JOPS@Standard.

MySQL Pricing: https://shop.mysql.com/enterprise/
Oracle Price Book: http://www.oracle.com/corporate/pricing/eplext.pdf
Red Hat Server pricing: https://www.redhat.com/apps/store/server/
Sun OpenSolaris, GlassFish Enterprise Server pricing: http://tinyurl.com/SunSoftwareVolume

Tuesday Dec 16, 2008

When you hit them and they smile, you know you did something right

I love to watch boxing. Big. Fan.  That's why when I read Sacha's post on open source business models - which questioned Sun's business model around GlassFish -  the first thing that came to mind was a boxing analogy. When you hit someone hard and they smile as if to say "You didn't hurt me", well, you've hurt them.

Sticking with the boxing analogy, Sacha's watched one fight tape in a GlassFish career that is racking up many wins. His post reflects having incomplete information. It's true, by the way, that Sun offers unlimited deployments of GlassFish for $25,000 - or should I say starting at $25,000 for organizations under 1000 employees.  When replaying the rest of the tapes, one can see that we have a tiered approach not all that different from the Java Enterprise System. The primary differences are that we group employees into tiers (to align with MySQL Enterprise Unlimited), and that the offering is for a single product instead of a multitude of products. Of course, GlassFish & MySQL Enterprise Unlimited makes a nice one-two punch. As for software infrastructure @ Sun, it's grown year over year.

Based on competitive intelligence, I know that JBoss offers site licenses as well, but it's unclear to me if JBoss/Red Hat has part numbers and dollar figures committed for an all-you-can-eat subscription. I can say that Sun puts a stake in the ground and assign a part number and dollar figure to an unlimited subscription - a commitment to the customer up front. How much are \*you\* paying for a JBoss Enterprise Application Platform site license?

Now to address Sacha's free training point, and to extend the boxing analogy well beyond its usefulness and probably effectiveness - Don King. One of the roles of a fight promoter is to drive attendance and revenue.  Regarding free training, it's a marketing promotion. Marketing has a budget, and we spend it as wisely as possible to drive maximum awareness and product adoption. Sporting events often take this approach to drive attendance.  When I attended a Chivas game recently, I wasn't too concerned about their long term viability when I got a free hat. Delivering free online training in a promotional context isn't going to break the bank. The content had already been created, after all. In fact, thanks Sacha for linking to our free glassfish training offer. As for any revenue "lost" to free training, we don't think quite that two-dimensionally. We're more than making it up when customers learn of the the strong value proposition of running GlassFish Enterprise Server on CoolThreads servers (a billion dollar business in-and-of-itself), and backed by the newly introduced Sun StorEdge 7000 Unified Storage system.

GlassFish adoption is through the roof, both in downloads and registrations. I personally like to take a look at the GlassFish GeoMap upon occasion as a reward for all the hard work that has been put forth by Sun and the GlassFish community.  I can emphatically say that this success is being followed up with a rapidly growing number of GlassFish Enterprise Server subscribers who value the support and service Sun provides.

To sum it up, when I see a blog post ending with "Sun being desperate" and "hail mary", I see a competitor trying to instill Fear, Uncertainty, Doubt - FUD.  That means the competition is smiling (errr, smarting) after they've been hit pretty good.  Me? I stay away from FUD and prefer to stick to the facts.

Wednesday Dec 10, 2008

Congratulations to the OpenSolaris team

As a side effect of being product manager for GlassFish, I've had to give up quite a bit of hands-on Solaris time. However, I've had the time to get OpenSolaris 2008.11 up-and-running in VirtualBox. Good stuff, and great job! I really look forward to trying out Time Slider. In the back of my mind, I'm wondering how this can be utilized for GlassFish configuration version control. In fact, me-thinks I'll try it out.

Saturday Dec 06, 2008

GlassFish Clustering in Under 10 Minutes

As we add features to GlassFish over time, one characteristic we do not want to compromise is ease-of-use. With this screencast I wanted to show how easy it is to set up a GlassFish high availability cluster.  While other clustering screencasts exist, I wanted to take a different approach. How long would it take, from download to deploying an app, to get a GlassFish high availability cluster up and running? This screencast does just that. While I could have gone faster - faster connection, eliminate narration, etc - I wanted to make it more typical and meaningful for viewers. Here's a summary of the timing:

  • Download: 2 minutes and 30 seconds. I cover this time by describing bringing up a slide and describing the clustering architecture
  • Installation and cluster setup: 4 minutes, 30 seconds.
  • Verifying and testing cluster by deploying a sample application: 3 minutes

Note, the embedded video is via the Sun Learning Exchange. This video is also posted to YouTube.

Update: You can play these video's in full screen. On YouTube there is a "Watch in High Quality" link just under the video (to the right) as well.

Friday Aug 08, 2008

GlassFish at Americas Sales Meeting

I've been in the Washington D.C. area the past few days extolling the virtues of GlassFish to the America's sales force. The good news is that the GlassFish session, an 8:00am affair, was packed. I must say that I was pretty darn surprised. Sales reps have \*dozens\* of products they can focus on, yet GlassFish drove a full house (60-70). Here's why:

  • Customer pain.  These days we often hear the phrase "pain at the pump" to describe how we feel about rising fuel prices. It's no different for how customers feel about how much they are paying for license fees at a time when the economy is in question. Now is not the time to raise prices, it's time to be more responsive to customer pain.
  • MySQL synergy. The business momentum behind the MySQL acquisition (inside and outside of Sun) is quite amazing. When you have a database, something needs to CRUD the data (I didn't know CRUD could be a verb either). When you have an application server, data needs to go somewhere. GlassFish and MySQL provide quite the one-two punch, without hitting below the financial belt.
  • Open Source - as in Freedom. Not just financial freedom, but the elimination of vendor lock-in. Customers feel beholden to closed-source products.  Customers do not feel beholden to open source communities. Buy only when you need to. Eliminate vendor lock-in. Sales now realizes that open source is a facilitator, not a competitor, because customers still want 24x7 support to reduce risk and downtime.
  • Transparency - Transparency facilitates trust. The GlassFish community is about as transparent as I've seen a community be. Sales enters an environment where a we have already established some level of trust and credibility. Their role is to extend that trust into a win-win business relationship.

Saturday Jul 26, 2008

Has Open Source killed the RFP?

When I was in sales, I participated in numerous Request for Proposal (RFP) efforts. To put it in perspective, I spent most of my sales career selling hardware to the government vertical (albeit as a software specialist). Governments are big on RFPs (impartiality). I also helped the commercial sales teams with RFPs upon occasion, so my peers suffered the same fate.

Product managers are not immune from RFPs. We tend to be back line support for those questions that need a bit more insight into the product (ex: roadmaps). Since being in product management I've noticed I'm involved in fewer RFPs. I can say with authority it's not because of a lack of commercial interest. GlassFish is doing quite well in that regard.

Customers are leveraging open source to reduce cost and eliminate vendor lock-in (among other things). Before the move, customers are still comparing, in the middleware space for example, GlassFish, JBoss, Geronimo, and Tomcat. I occasionally get comparison spreadsheets from potential customers. However, the process of the formal RFP seems to be pushed aside. Here are some of my thoughts as to why:

  • Transparency. Open source is by definition "open". A vast majority of the information that is typically requested is publicly available.
  • Google. The information that is public available is easily found.
  • Community. The information that is not available can be quickly addressed by posting an email.
  • Resources. Organizations do not like the RFP process. It's time consuming and people-consuming. Thank goodness because vendors do not like RFPs either - it is equally time consuming :-)
  • Viral growth. Open source technologies often get deployed first at the departmental level. They tell two departments, who then tell two departments, and so on, and so on. Before you know it, the open source technology is already the company standard.
  • Value Proposition. The value proposition of using open source in general, or a specific technology in particular, is often compelling enough to "sell itself". For example, GlassFish offers enterprise features, is 1/10th the cost of its closed-source peers, and outperforms them as well.

The blog title truly is a question, not a statement. I've thrown out some supporting thoughts, but I don't want to claim an authoritative statement. Regardless, I know it is not an absolute truth. So the question remains, has open source generally killed the RFP?

Tuesday Jul 22, 2008

This is not an ALERT

In an effort to gain visibility in a world of information overload, the media invented the "ALERT". However, there seems to be no consensus as to what an "ALERT" means. The result is "The Boy Who Cried Wolf" syndrome, where an ALERT more often than not adds to the noise.

I made the mistake of subscribing to ZDNet mailing Open Source ALERT mailing list. Note to ZDNet, a post by Dana Blankenhorn is not worthy of an alert. That is not to discredit Dana whatsoever. I subscribe to Dana's blog. However, a daily blog post is not ALERT-worthy. I unsubscribed. ZDNet lost a communication channel to me, their customer.

Note that cable news is not immune, where they spend endless hours of analyzing a single topic. An ALERT is intended to notify the viewer of something eventful, although it rarely does so.

My thoughts on good vs bad alerts.

Good alerts - including mandatory Yahoo, Microsoft, Google, Apple references:

  • Yahoo buys Microsoft
    Microsoft buying Yahoo lost its alert-ability months ago
  • Google Trends shows googling Google on the decline
    Not that it will ever happen
  • Microsoft places [anything significant] under an open source license
  • Apple opens up iPhone to developers, and AT&T Unlimited data plan is really unlimited, enabling SlingBox owners to Sling.
    GlassFish & MySQL Unlimited really is Unlimited.
  • Tsunami. Earthquake.

Bad alerts:

  • Microsoft buys Yahoo
    See "endless hours of analyzing a single topic" above
  • new iPhone model released
    "... endless hours ..." yada, yada, yada
  • Anything about Britany Spears
    Alerts should be newsworthy
  • Alien Invasion
    They're already here

Friday Jun 27, 2008

Save big bucks with GlassFish & MySQL Unlimited

While other vendors are raising their prices, Sun decided to do just the opposite. We decided to save the customer money. In fact, not only can customers save money now, going forward they can scale their business without having to scale their budget. Introducing GlassFish & MySQL Unlimited, an offering that combines two fast, reliable, and very popular products, based on open source, into a single subcription offering.

Starting at $65,000, a customer can deploy as much GlassFish Enterprise Server and MySQL Enterprise Server as needed to meet business requirements. No counting CPUs. No counting cores. No audits or true-ups. No counting servers. And you don't have to "go back to the well" and ask for more money.

Do the Oracle math. Don't forget to double the price-per-socket for quad-core x86 servers and for 8-core Niagara servers. And add support.

Do the Sun math. For organizations up to 1,000 employees, it's a $65,000/year for unlimited use. If you want 24x7x365 support for GlassFish Enterprise Server (for a more fair comparison), it's $80,000/year. No worries, the MySQL component already includes 24x7x365 support among other things. The offering extends beyond 1,000 employees, but you get the point.

Do the math using your budget. Just think of what you can do with what you save! If you can't think of what you can do with the savings, rumor has it that Java CAPS 6 is a great deal :-)

Note: Mark has some math as well.

Sunday May 11, 2008

GlassFish in review @ JavaOne 2008

It's been nearly 5 months since I've last blogs. Clearly I'm in the realm of the non-blogging heathen. No excuses. Mea culpa.

The good news is that the GlassFish community has been plugging away at GlassFish with substantial progress.  There have been hiccups along the way and surely there are more to come, but overall GlassFish is forging ahead nicely.  Here's a summary of GlassFish happenings at JavaOne 2008.

  •  Product name change. "Sun Java System Application Server" was renamed to "Sun GlassFish Enterprise Server".  The new name inherently describes the relationship between open source product and the Sun product. Simply stated, Sun offers support of GlassFish through the Sun GlassFish Enterprise Server.
  • GlassFish Enterprise Server Unlimited. Deploy as much GlassFish Enterprise Server as you want, starting at $25K. That's less than deploying some competitive application servers to a single T5140. I don't have the web page up for the Unlimited offering yet. This is one of those afore-mentioned hiccups.
  • Partner Initiative for Sun Partner Advantage program. Over 40 GlassFish partners! Check out the partner showcase. Want to become a GlassFish partner? Here's how.
  • GlassFish v3 Technology Preview 2. You may recall Technology Preview 1 (TP1) at last year's JavaOne. Since then we have been busy delivering multiple releases of GlassFish v2. Not resting on our GlassFish v3 laurels, Technology Preview 2 delivers the following over last year's Technology Preview 1. I'm sure I'm forgetting something but I'll chalk it up to being way past my bedtime :-)
    • Support for OSGi
    • Dynamically extend GlassFish v3 by adding modules to the lib directory.
    • Java Persistence API, JDBC Connection Pooling
    • File & JDBC Realm support, SSL support
    • Complete asadmin command line interface support.
    • NetBeans support, with GlassFish v3 available in NetBeans update center
    • Update Center support with Solaris IPS packaging. The following components are available using the update center:
      • Web administration console (subset of GlassFish v2 console)
      • Partial EJB 3.1 support (stateless session beans w/local interface)
      • EclipseLink
      • jMaki, JSF, Woodstock
      • JRuby runtime, Grails support
      • Project Jersey REST apis (JAX-RS)
      • Metro Web Services stack
  • Project WebSynergy. Bringing together the best of LifeRay and OpenPortal with GlassFish. Check out The Aquarium writeup.
  • GlassFish ESB. Read all about it.
  • Sun GlassFish Communications Server. Sun's SIP-enabled application server based on Project SailFin. SailFin, in turn, is based on GlassFish.
  • GlassFish Unconference. Arrived late due to a delayed flight. However, I had the pleasure to lead the "Production & Performance" discussion. Learned a lot about how folks are deploying GlassFish. Looks like we can improve NetBeans integration by addressing a bug or two. The good news is that we received multiple kudos for GlassFish ease-of-use and for call flow monitoring.
  • GlassFish Party. I can neither confirm nor deny I over-consumed alcoholic beverages. Hiccup (unrelated to afore-mentioned hiccups).
  • GlassFish BOF. Picked up some good feedback at the BOF on how to improve GlassFish.
Believe it or not, I only had the chance to attend one session the entire conference - the opening general session. Had to go back to UStream to watch GlassFish v3 start global thermonuclear war during Bob Brewin's afternoon keynote. The rest of the time was dedicated to talking to partners, customers and long-time friends.

Friday Dec 21, 2007

lzPack GlassFish installer on the Mac

First off, I'd like to give a big thanks to Julien Ponge for the GlassFish V2UR1 lzPack powered installer. Impressive. Works like a charm. I thought I'd blog about the Mac installer. The installer defaults to installation in the /Applications directory. If you choose to do this, odds are you will run into problems with not being able to create the domain directory during the configuration step due to a lack of permissions to write to the /Applications/GlassFish directory.

There are a couple of options.


  1. Continue installation in /Applications directory. The setup fails to create the domain, but you can create the domain after the fact and point to a user-writable directory as follows:

    # /Applications/GlassFish/bin/asadmin create-domain --domaindir /Users/jclingan/domains --adminport 4848 mydomain
    # /Applications/GlassFish/bin/asadmin start-domain --domaindir /Users/jclingan/domains

    Note /Users/jclingan/domains is a top-level directory for GlassFish domains, so a mydomain subdirectory will be placed under that directory. Most folks don't know about the --domaindir option so I thought this option might be useful to cover.

  2. Select a different install directory. Simply choose a subdirectory off of your home directory, such as /Users/jclingan/GlassFish. This is similar to what you would do with the jar installer and the permissions issue goes away entirely, resulting in a clean install.

Thanks, Julien, for the effort. A very welcome addition!

Wednesday Dec 19, 2007

GlassFish V2 Update Release 1 Released

Three months after the release of GlassFish V2, we already have an update release. Alexis has a good writeup of what's in this release. I'd like to point out a few items in a bit more detail.

First, the Update Center is getting more visibility. It's a great GlassFish feature that many developers do not even know exists! A growing number packages are available and plug right into GlassFish. We are also trying to grow the number of packages as well, both with community input or by the community itself. Example: Open ESB is right there, click, click, restart. Your ready. Or Try jMaki (no restart required).




Next, the support center offers up-to-date information of all kinds, aggregating a ton of information into one convenient location. In particular, I like the feed showing the list of bugs fixed. It should be noted that this feature does require product registration. On a related note,  many bugs have been fixed. For those of you waiting for an update release before you deploy, well, here it is.






Thursday Sep 27, 2007

BEA taking notice of GlassFish

Looks like BEA is taking notice of GlassFish. Bill Roth offered some counter-arguments to our "10% better performance than BEA WebLogic 9.x" comparison, which is both expected and an appropriate thing to do. Bill asked some open-ended questions which I addressed for the benefit of the BEA community. I also offered some clarification around some of the points that Bill made so the WebLogic community has a better understanding of the GlassFish community. Bill's a busy guy (moderated comments) and the blogosphere is a moving window of conversation. In fear of missing the that window, here is a copy of the comment I submitted to Bill's blog.

Update: Alexis has a view on subject as well.
Update: Thorleif chose GlassFish over WebLogic (2nd comment) for the very reasons stated below.
Update: Hmmm, Bill never published my original comment to his blog, only the courtesy FYI ...
Update: Bill's SPAM catcher got in the way (see comments). No worries Bill, my email filter occasionally junk-mails my own emails :-)


Bill, GlassFish V2 and the Sun Java System Application Server 9.1 are essentially the same, with the latter applying the Sun brand, indemnification and support.  Sun benchmarked the Sun Java System Application Server 9.1 instead of GlassFish V2 because SPEC rules require a supported product. I highly recommend that instead of asking open-ended questions, you engage the GlassFish community via the user forums or the mailing lists where these and future questions can be easily addressed. To be completely transparent, I've updated the GlassFish User FAQ with the differences (GUI installer, 3rd party JDBC drivers, HADB) between GlassFish and the Sun Java System Application Server. These are not "significant differences" in our view.

Regarding our JVM, we don't need to be "pressed", simply asked (refer to the previous comment on user forum and mailing list). Yes, there are performance improvements between Java SE 5 and Java SE 6. However, since neither BEA WebLogic 9 nor BEA WebLogic 10 support Java SE 6, WebLogic customers cannot benefit from these performance improvements. It is good to hear that BEA has improved WebLogic tuning since the last submission since that benefits the larger Java community. The GlassFish community has made significant performance improvements between GlassFish V1 and GlassFish V2. The GlassFish community continues to actively work to improve performance across the board in a variety of areas.

I'll concede that benchmarking is a game of leapfrog. The main point is that GlassFish is performance competitive and extremely feature/price competitive at $4,500 for 4 sockets, which is why Gartner, Forrester, and Current Analysis are taking notice.


John Clingan
Sun Application Server Product Manager

Saturday Sep 22, 2007

GlassFish V2: The start of something great

What a difference two weeks make. Two weeks ago I was pulling (what's left of) my hair out trying to get the T's crossed and the i's dotted on GlassFish before the Monday launch. Pre-and-post-launch I've have been talking to potential customers, the sales force, analysts and the press. The word is getting out and the feedback is extremely positive.

I can't help feel that this GlassFish launch is the start of something great. Yes, Sun has had an application server (in one form or another) for almost a decade. Yes, GlassFish V1 has been around for over a year - even in production. But it all seems to have come together with GlassFish V2. Grrreat performance. New cost-efficient subscriptions for support. Enterprise ready. Open source community. Fully Java EE 5 compatible. The value proposition is looking good.  Not to put ASCII in their HTML, but Gartner seems to agree.

As we, the community, move forward, I'll do my best to keep everyone abreast of GlassFish momentum. The message may not always come from The Clingan Zone, but I've got prolific friends :-)

Monday Sep 17, 2007

GlassFish V2 released with cost-efficient annual subscriptions

There is a good reason for the rather long pause at The Clingan Zone. As mentioned previously, I moved out of the field and into corporate as the GlassFish Group Product Manager. In that role, I have been working in the background doing my part to help move GlassFish V2 towards the goal line. As of today, we've crossed that line. GlassFish V2 is formally released!!

GlassFish V1, the first Java EE 5 application server available, focused on developers with Java EE 5 ease-of-development features, low resource consumption, dynamic resource configuration, etc.  GlassFish V2 adds out-of-the-box enterprise features. In particular:

  • Clustering - Cluster for scalability high-availability. New to GlassFish V2 is in-memory session replication for high availability; robust and easy to set up. For 5 9's of availability Sun continues to offer HADB.
  • Advanced administration - From a centralized GUI console (or  CLI) users can manage and monitor the clusters and applications. Clusters can be created and grown dynamically to meet user demand.
  • Best-in-class performance - GlassFish V2 offers great performance and is the fastest open source application server available. In fact, GlassFish V2 is 10% faster than BEA WebLogic 9.x on the same hardware (compare here and here).
  • Microsoft .NET 3.0 Web services Interoperability - At it's core,  Project Metro offers industry-leading web services performance with JAX-WS. In addition, Sun (through Project Metro) and Microsoft have worked together to ensure secure, reliable, transactional and high performance web services interoperability between Java EE and Microsoft .NET 3.0.

Enterprise features do not compromise ease-of-use. To date the feedback we have received on ease-of-use has been overwhelmingly positive, especially for clustered deployments. Ease-of-use will be a top-level theme for GlassFish V3 (more on that in a later post).

Sun's commercially available counterpart to GlassFish V2, the Sun Java System Application Server 9.1, is offered with new annual subscriptions for support, including live transfer for issues impacting production availability. In fact, we have dropped prices by up to 75% to be in line with other open source offerings. Yep, you can buy online. Note, GlassFish V2 is at feature parity with the Sun Java System Application Server 9.1 - we're not penalizing the open source community for committing to open source bits.  However, if production support is required to reduce the impact of production issues, we have your back. With great out-of-the-box enterprise features and cost-efficient pricing, the Sun Java System Application Server 9.1 offer a great value proposition whether you are an enterprise looking to drive down costs or a startup looking for a low barrier to entry.

Download GlassFish V2 or Sun's commercialized counterpart, the Sun Java System Application Server 9.1. Take it for a spin. Have questions? Check the documentation or the FAQ on how to get help.


John Clingan-Oracle


« July 2016