Sounds like some hogwash Satish. I don't hate JBoss, and I prefer it to just about anything else out there, but you force me to point out the flaws in these talking points which read more like an ad than an honest analysis.
1) JBoss may have a bigger and more active community base, but Glassfish appears to be eroding it largely at the expense of JBoss (and Weblogic, shocking huh?)
See this link, where it shows Glassfish overtaking JBoss, and skyrocketing in popularity.
https://www.sun.com/offers/thankyou/thankyou.xml?id=glassfish_jboss.pdf&original-page=/offers/details/glassfish_jboss.xml
Glassfish was found to be the app server of choice for 73% of new JEE projects according to Ohloh. (WOW!)
2) ? Like what? And does it really matter? Performance, scalability, usability, and reliability all trump non-standard niche features.
3) Dependence is not as high on Red Hat? Meh, this is a weak one. First, less vendor lock-in is the main benefit of OSS--and both are OSS, so this isn't a major issue even if the dependence is slightly greater. Also, most large conservative corporations would kick you out of the room if you suggested going without support on EITHER JBoss or Glassfish, so either way you're attached to someone. At best this is a modest JBoss win for small shops where they're willing to go without commercial support.
4) Oracle has made it clear that Glassfish is not "going away". And they couldn't kill it if they wanted to at this point. It would likely just get forked- that's the nice thing about OSS, you should know better than to suggest that Oracle would bury GF just because it might cut into Weblogic's market share. WebSphere is a more likely competitor for Weblogic anyway, not Glassfish. Oracle's not stupid enough to risk losing the cash flow of Glassfish support (by burying it) AND having a forked version (or other competitors) still end up competing with Weblogic in the long run anyway. It's more in their interest to keep supporting it as the replacement to their failed OAS/OC4J junk platform.
5) JBoss is slightly more developed for now, but that era is coming to an end pretty quick. Glassfish is the new favorite JEE server for developers everywhere I look. And GF does very well on benchmarks. JBoss usually does poorly. Perhaps the JBoss code base is starting to suffer from the inevitable bloat that tends to infest software shortly before it starts losing popularity? Or maybe it was just inefficient from the start, who knows. I do know that Glassfish is fast based on personal experience as well as benchmarks.
6) Community was already discussed... and apparently JBoss' community is losing ground every year.
7) This is a good point, and I hate super-vendors as much as the next guy, but Oracle has a fairly impressive track record when it comes to competition and they're going to be tough to stop. Not that I want one to "win" over the other--I'd rather have more than one implementation! But Glassfish has the "soul" of Sun Microsystems and it'll probably be some number of years before the usual Oracle bloat ruins that (if it does), but by then at the rate JBoss seems to be going, it'll be even more bloated and aged.
Based on the graphs in the link I posted, Glassfish is stealing the show (and market). Weblogic AND JBoss are on VERY obvious, steady, and long running downward trends which can't be denied in those stats. The IBM scum's overpriced proprietary WebSphere bloatware is flatlined, and will probably remain on life support indefinitely as long as we have moronic corporations that refuse to adopt OSS alternatives simply because it's OSS.