EnterpriseDB’s Postgres Plus Cloud Database product is now in beta. It is a great product and is one of the easiest ways to get started with Postgres or with a relational database in general. Yes, as an employee of EnterpriseDB I am extremely biased. However, I tell you the truth, speaking as somebody who works with people who are deploying Postgres and Postgres Plus Advanced Server day and night… this product is great. For now I will only tell you one of the many things I like about this product.

Typically when people think of cloud computing they are drawn to the concept of elasticity. More specifically, the ability to apply hardware and software resources to a task as needed.  For instance, the software I am running is saturating the CPUs of the server it is running on. In a cloud environment, the software can detect this and brings more CPUs or more servers to handle the workload. When the CPUs resource consumption returns to normal levels the software can return the CPUs and servers to the cloud making them available for other tasks running in the cloud.=

While Postgres Plus Cloud Database has elastic properties, what I like the most about it, has nothing to do with elasticity. It is how quickly and easily multi-host PostgreSQL clusters can be provisioned. In a matter on minutes, I can set up a Postgres cluster with several read replicas and automatic failover. I can them immediately begin testing my application against this set up. I don’t need to find servers (or order them),  install 1..N versions of Postgres, install PGPool, configure and tune Postgres, configure and tune PGPool. It can all be done in a matter of minutes.  Even if you don’t intend to deploy in the cloud, you can begin testing the database deployment which  you have designed immediately.  You don’t need to wait for your servers to arrive and go through ½ a dozen vi sessions. You can do it now!

I have told you about one of my favorite features of Postgres Plus Cloud Database. I will tell you about others later. However, there is one other thing I want to mention in this post, since Postgres Plus Cloud Database will enable you to provision both multi-host PostgreSQL and Postgres Plus Advanced Server clusters in under 10 minutes, you can setup a multi-host Advanced Server deployment and subsequently, via our migration tools, migrate yourself off Oracle and into the cloud. All of a sudden you have opened up new possibilities for your enterprise.  To get more information about Postgres Plus Cloud Database go to this URL: http://www.enterprisedb.com/products-services-training/products-overview/postgres-plus-cloud-database

If you're new here, you may want to subscribe to our RSS feed or follow us on Twitter for product announcements, event information, and industry news.

Bruce Momjian recently did a blog on Postgres Monitoring tools. I want to point out a unique and useful feature of Postgres Enterprise Manager from EnterpriseDB.

I frequently hear from prospects that they would like us to help them debug periodic performance drop-off their PostgreSQL server is encountering. When we get called, usually the opening description is something like this:

“Yesterday between 3 and 4 pm my users all started complaining that the system was incredibly slow. At roughly 3:45 pm a few disgruntled users showed up at my door with pitchforks. Thank heavens performance reverted to normal levels around 4:05 pm.”

The good news is that performance is now back to an acceptable level. The bad news is nobody has an inkling why performance fell off a cliff. In addition, the DBA or whoever is deemed to be responsible for making sure this sort of thing does not happen, is now living in fear of the next comparable event on the database server and the pitchfork wielding users will return in greater numbers and with even less patience. Yes, some PostgreSQL and Linux experts can go in and evaluate a system and gain a few clues and periodically make a reasonable diagnosis about why database performance suffered. However, what happens more often than not is that a number of probes are installed on the system to gather more data for the purpose of doing a better analysis of the next performance drop off. Getting to the root cause of such spikes can take many iterations of installing probes and subsequently analyzing the collected data. To impede progress further, some where in the days or weeks that this cycle may require to debug the issue, a second server starts encountering sporadic performance problems. Now the resources to resolve the issues are spread thinner and oh dear, what if they are two entirely unrelated problems!

The source of such performance problems may be a variety of things, here are some examples:

1) Postgres checkpoint settings are not tuned properly
2) The operating system starts swapping as a result of some activity the database is doing.
3) Many users coming on line at roughly the same time and start using the application because the CEO sent out a corporate wide e-mail telling people to take action on something.
4) A report is being run against the database periodically
5) Vacuum settings are not properly tuned.
6) Backup processes or other external processes running on the system.
7) The database has stopped doing index scans on a large table and is now doing sequential scans

To determine which things from the above list is the problem or to determine which of the many other possibilities that may be happening, database and system probe data is required. Further, graphical analysis of this data may be required to gain understanding of the problem.

If Postgres Enterprise Manager (PEM) is installed some place on the network, and a PEM agent running on the data server where the performance drop off occurred you have everything you need to diagnose and trouble shoot the problem. You don’t need to wait for the problem to repeat itself and hope you are able to gather enough data.

1) You have the database and system probe data.
2) You have the graphical tools to find the smoking gun in the data that will lead you to the root cause of the problem.

The specific tool within Enterprise Manager that will guide you to the problem is called the Capacity Manager. It repeatedly gathers statistics about what is happening on the system to allow you to forecast “what will happen on the system”. However, what is often more interesting is looking at the statistics to find out “what the heck did happen” on the system.

I have all ready done a video on how to use the Capacity Manager to forecast disk space needs. You can find it here: http://tinyurl.com/7sd3syx

In the coming weeks I will be creating several additional online demos for Postgres Enterprise Manager. The first of which will be how to use Capacity Manager to debug a sporadic dip in performance.

So please stay tuned…

EnterpriseDB was a Platnium Sponsor of TheServerSide Java Symposium Conference last week.  Here are a few key takeaways and highlights from the conference:

- It was great to be at a Java conference that was entirely based on Server-side Java. Server-side Java is one of my few technical passions (as I mentioned in previous posts, PostgreSQL has become another).

-James Gosling gave a great keynote. It had been difficult for me to not see him at JavaOne this past fall. I loved seeing him together with the Java crowd again. His keynote was very forward looking as he discussed his thoughts on the industry particularly around cloud computing. James was entertaining and insightful as ever.

-There was a panel specifically about the JCP. Patrick Curran, Reza Raman and James Gosling were the panelist. Cameron McKenzie,from TheServerSide.com was the show’s MC and moderator. For the most part, tough questions were asked and answered. At the end of the day, I felt as though decisions had been made, Java and the JCP were moving forward again. All though it is now with out some of the previous industry stalwarts such as Doug Lea. A quick check of the TIOBE index shows that Java is in good shape. It will be interesting to see how Java EE and Java SE progress over the next year. I must confess, I am disappointed to not see more Java EE 6 certifications at this point. However, I was pleasantly surprised to hear that Resin has passed the Web profile TCK.

-In terms of conference themes, cloud is definitely all the rage this year. It was prevalent in every keynote and had it’s own dedicated track.

-Oracle was also a Platinum sponsor at the event. Jerome Dochez, the GlassFish architect gave several good talks on GlassFish, including the vendor sponsored keynote. As usual, he was great. I got the overall impression that GlassFish was in great shape, which naturally made me feel good. In the coming weeks, I plan to write another blog on how to get it working quickly and easily with PostgreSQL.

-I spent time at the Resin booth. This is a product I am definitely going to check out. Many people at the conference told me it is a great application server.

-Some vendors who were notably missing from the conference were JBoss and Cloudbees. This was disappointing because I wanted to hear more about their product plans and what they are doing these days. Although Cloudbees did not attend the conference, they got a lot of love from James Gosling in his keynote. Hudson (Kohsuke Kawaguchi the creator of Hudson works for Cloudbees) also got love from Rod Johnson in his track keynote on cloud computing.

-I attended Rod Johnson’s track keynote on cloud computing. All though I disagreed strongly with some of the things he said, I genuinely enjoyed the keynote. Code2Cloud also looks rather interesting. As for what I disagreed with, Rod claimed that there were three platforms that would be relevant in the cloud: “Spring, Ruby on Rails and .NET”. I will give him .NET. I will even give him Spring as being relevant in the cloud. As for Ruby on Rails, every indication I can see is that it has a small and passionate community. However, it is no longer seeing the growth it saw several years ago. If you believe in the TIOBE Index as a status of how programming languages are doing, Ruby has seen a decline year over year. Also, missing from his list was Java EE. With the release of Java EE 6,  the complexity issues that led to the creation of Spring no longer exist. Java EE 7 if executed with pace will definitely be a very relevant platform for cloud computing.

So, on to what I was there to do:

EnterpriseDB gave me the pleasure of delivering a vendor technical session. My talk was entitled “PostgreSQL from Development to Deployment.” It is the pre-cursor for a talk Bruce Momjian and Robin Schumacher will be giving at MySQL Con in less than a month.  I felt very comfortable with the audience at the show since I have been working either on or with Server-side Java products and technologies ever since they appeared.

Several days before the conference began, much to my disappointment it became clear that Karen Tegan Padir would not be able to deliver the vendor keynote. As a company, we made the decision to have me present the keynote. I was a bit nervous going into this given it has been a years since I had spoken to an audience of this size. Karen and I spent some time discussing what I should talk about. From there, I spent several nights reflecting on my time at Sun building the Java Platform; what went well and what did not. For those of you who don’t know, I was the first Java Consultant / ISV Engineer hired by Sun’s Javasoft division and was part of the original Java management and architecture teams for the J2EE platform. In the keynote, I described what went well and what did not go well in my nearly 15 year adventure with Java. I explained the similarities I am now seeing in the development of the cloud platform and how the lessons we learned developing the Java will apply to the cloud.

From what I could tell, the keynote was well received. Several folks came up to me and said it was a great presentation for this audience. I had plenty of friends in the audience who would have been more than willing to tell me if it was a dud. I actually got a laugh (at a joke!) during the presentation. However, the biggest indication that it was a success was that my technical session on PostgreSQL that followed the keynote was extremely well attended. We had about 54 people show up. Not bad given that I would estimate the overall conference attendance to be about 350 people.

There was also plenty of interest in PostgreSQL especially after I explained the history of the source code. A lot of folks came by our booth to talk about it further. It took me back to 1996 when I first started talking to folks about Java. Although PostgreSQL has been around for a very long time, I feel as though many people are just starting to discover it as though it were brand new. For those of you who attended the session, I will be writing a blog on a how to set up the demo.

In summary, it was great to spend time with old colleagues and friends. It was also great to make a bunch of new friends and hear what people have been up to in the world of Java. However, it was tough to lose so badly at blackjack!

A while back I talked about my experience at Java One and mentioned Thomas Kurian’s all business Keynote. It was intermixed with technical depth on topics such as Contextual Dependency Injection in the latest release of Java EE and the 3D graphics now available in JavaFX.

Yes, it was a good Keynote. However, it made me think back to the greatest Keynotes I have ever seen at any technical conference. I have seen roughly forty KeyNotes at technical conferences. With out question, the top five were all delivered by Scott McNealy. So what so great about them?

When Scott gets on stage and starts talking, you get the impression that he would do his job for free. It appears he is thinking “I can’t believe I get paid to do this”. He is having fun, and so is everybody in the audience. Arriving in a leather jacket, sneakers and a smile that never stops, joke after joke intermixed with excellent insights about the IT industry. He came to work today to have fun. Or his own words “to kick butt and have fun”. You can count on everything that comes out of his mouth falling into one of three categories:

1) Very funny. “Microsoft Monopoly: Where every Bill has a Bill on it”.

2) Very insightful. “The Network is the Computer” Perhaps a phrase that is out of date in the early 90’s it was exactly where the industry needed to head and was heading.

3) Something many are thinking and nobody wants to say. “Greenspan is worrying about the economy growing to fast? And that is bad why?” OK, I had a better example but I just don’t have his gumption to repeat it.

The trade rags would be full of his quotes for weeks after he spoke. I don’t think PG West will be any exception.  In the later years of Sun’s existence,  his tone mellowed a bit (just a bit). However, now that he has been unburdened by being an officer at a company that was hurting and was for sale, I expect him to return to his old form and then some.

Scott was always somebody who believed that if you innovate properly everything else will take care of itself. For decades this formula worked and he built one great company where many of the world’s top engineers loved to work.

This is a very exciting time for PostgreSQL. I don’t think there is any doubt, given what has happened in the market recently,  it will become the most widely used open source database. It is a pleasure to work with and it’s momentum is building. Again, it feels like the early days of Java where potential abounded. Having Scott speak at one of PostgreSQL’s most prominent events simply cements my opinion on the trajectory PostgreSQL is on.

For those of  you who would like learn a little more about Scott, one of the greatest Silicon Valley entrepreneurs of all time, I recommend the book “High Noon: The Inside Story of Scott McNealy and the rise of Sun Microsystems”, http://tinyurl.com/327zb5b. Many fun stories about bold business moves.

Kudos to Karen Padir for bringing PostgreSQL and Scott together. Should be a great conference!

Thoughts on JavaOne.

I attended JavaOne 2010. I attended the first JavaOne in 1996 and have only missed three over the past fourteen years.

This one was extremely different for several obvious reasons:

1) It ran concurrently with Oracle Open World
2) It was hosted by Oracle and not Sun

Very different. Here are the things that struck me the most:

Thomas Kurian in a suit and tie delivering a polished presentation but all business. All the demos were great but a few were delivered via a video recording. Made me think back to Rich Green wearing jeans, boots and a black tee shirt also giving a polished presentation but cracking a few jokes giving cutting edge demos some of which would crash, but crash with humor and charm. Kudos to Nandini who was on stage with both of them and adopted beautifully to both styles.

Most of my time was spent in the Hilton this year. Not the Moscone as in years past. It was impossible to get a feel for how well the show was attended. The one Keynote at Moscone was packed. Not sure how many were Oracle world attendees and how many were JavaOne attendees.

Some speakers that were notably missing:

1. James Gosling
2. Gavin King
3. Rod Johnson

However, plenty of great talks, here are some of the ones I attended that I especially enjoyed.

1. Technical KeyNote (I especially enjoyed Mark Reinhold’s use of animation in his slides, came off great). As always Roberto and Arun did a great job. Couldn’t stay for the third speaker.
2. Mark Little’s talk on Transactions given at the JBoss booth.
3. RDBMS vs. NoSQL a shootout. RDBMS guy vs. a Java guy. Two guys from Verisign guys did an amazing job. I will have a separate blog on this.
4. What’s new EJB by Mahesh and Marianne
5. Spring Roo

Also, some really exciting new comers with booths:

1. Cloud Bees. http://cloudbees.com/ A platform as a service company founded by Sacha Labourey. Also working with Sacha are Vivek Pandey and Ryan Campball. Three excellent people who I have had the great pleasure of working with. In my opinion a can’t miss company. I with them all the best!!
2. InfraDNA http://www.infradna.com/ . A company founded by Kohsuke Kawaguchi. He is joined by Kedar Mhaswade. Also, two of my favorite people and a can’t miss effort. Wishing them best as well.

All in all, Oracle gave a pretty good show. It is going to be a very interesting year for Java. However, not nearly as interesting as it will be for PostgreSQL . Seriously, after working with these folks and this community, I now feel about PostgreSQL in 2010 like I did about Java in 1996.

It has been years since I blogged. I started an external blog at Sun with good intentions. However, as an engineering Executive Director at Sun, and later a VP at Oracle I was never deep enough in the technology to contribute anything beyond what the many experts who worked for me were all ready blogging about.

What many Directors and VP’s who blog write about is industry trends. Frequently extolling features of their products over their competitors. There were many times at Sun I wanted to blog about others in the Enterprise Java space. However, as a result of being tasked with delivering Java EE 6, I had to be oh so careful to ensure we had the votes for those fragile JCP ballots.  It would be difficult to blog about why I thought Google was harming Java and to subsequently seek their vote for in the Java EE 6 community review. At Sun, when it came to a JCP ballot the stakes were always high for us. We had invested the most by far and had the most lose if it failed. The JCP was was often under fire for one reason or another, and I was not about to put gasoline on it

Well now, here I am on my new role working for an east coast company:  EnterpriseDB. I am the VP of professional services. It is a great company and I am very happy to be here.

In general terms, EnterpriseDB is a company built around PostgreSQL. We make one click installers for PostgreSQL. This is one of the many things we do for the community. Some of the other things include employing key contributors to PostgreSQL and providing them with freedom and time to work in the community on important issues and evangelize PostgreSQL.

We offer support and many other professional services for PostgreSQL. In addition, we offer a product called Postgres Plus Advanced Server. PPAS as we call it, looks and feels like PostgreSQL, however it offers features for advanced manageability, performance, ease of use. In addition, comes with an Oracle compatibility layer which enables people to migrate an application off Oracle to a lower cost subscription based solution.

Getting back to blogging:

I have wanted to blog for a while. I thought long and hard about what I should comment on regularly. I am not a PostgreSQL expert, all though I am learning fast. However, there will always be at least fifteen people who work for me who know it better than I do. Not to mention the many other employees at EnterpriseDB and the PostgreSQL community members.  I definitely want to talk about industry trends, especially why nobody should be using MySQL any longer.  However, what I mostly want to talk about is PostgreSQL and Java. Java developers typically want the datbabase out of their way. However, there are many great features in PostgreSQL that make it a joy to Java development with.  What are they? Well stay tuned to this blog and I will tell you. However, in the meantime, let me say, I have now been working with PostgreSQL for several months. I have created databases and I have built the source. It is a joy to work with and so are the people who are so committed to it’s success.

Please stay tuned for the following topics:

1. JavaOne 2010 report: How did Oracle at hosting their first JavaOne?

2. PostgreSQL 9.0 together with GlassFish

3. PostgreSQL and Hibernate

4. Cool Java tricks (java based stored procedures etc.) in PostgreSQL

5. Object Relational Mapping in PostgreSQL

6. Why you need to get off MySQL and on to PostgreSQL