New-age Transactional Systems - Not Your Grandpa's OLTP
John Hugg discusses high volume transaction processing applications with high and low frequency profiles, and how VoltDB can be used for that purpose.
The content has been bookmarked!
There was an error bookmarking this content! Please retry.
Posted by Rob Thornton on Feb 16, 2007
IBM has updated a 2004 article on Java EE best practices, compiling a list of 19 practices. They range from always use MVC to prefer JSPs as your first choice of presentation technology.
The article is an updated version of one published in IBM’s WebSphere Developer Technical Journal in 2004. They’ve updated it to include current technology and they include practices that the authors consider standard but are not always followed. Number two on the list is: “Don’t reinvent the wheel. Use common, proven frameworks like Apache Struts, JavaServer Faces, and Eclipse RCP. Use proven patterns.” They expected frameworks like Struts to replace homegrown MVC frameworks, however they say:
What has proven surprising is that this has turned out to not be the case. We still see many companies maintaining or even developing new user-interface frameworks that are functionally equivalent to Struts or JSF. There are many reasons why this could be true: organizational inertia, “not invented here” syndrome, lack of perceived benefit in changing working code, or possibly even a slight sense of hubris in thinking that you could do things “better” than the open-source developers did in a particular framework.
However, the time is long-past when any of these reasons is worth using as an excuse not to adopt a standard framework.
The complete list of practices is:
A Guide to Branching and Merging Patterns
Getting Started with Stratos - an Open Source Cloud Platform
Improve Java Garbage Collection, Runtime Execution, and JVM visibility with Zing
18 agile and lean practices for effective software development governance
Reading this article made me feel disillusioned and a little bit angry. Much of the article, where the author sticks to non-product guidance like "layer your app" is good. But several of the technology-specific tips are misguided at best. For example:
1. It implies that the only viable view technologies are JSP and XML/XSL. What about the excellent Velocity and FreeMarker templating engines? They are great for preventing developers from embedding business and controller logic in views.
2. Use Struts? Maybe five years ago before Spring MVC came out. Spring's offering is so much more testable, extensible, and decoupled. I don't know much about Spring Web Flow, but it may be even better still, certainly for web workflows of any complexity.
3. Use Cactus to test J2EE components? I admit to doing this myself before seeing the light on POJOs. Write your application using POJOs, and you can test them using plain old JUnit without complicated and slow in-container testing. Then wrap them in a trivial EJB if you need to (although with a DI container like Spring you usually don't even need that).
4. Use J2EE security for authorisation? For most real-world apps, where access is per-entity based on complicated criteria, it's useless (ref: www.owasp.org/index.php/Access_Control_In_Your_...). JAAS is only good for authentication. So overall, you're better off using a pluggable security framework like Acegi for both these aspects of security (are there any real alternatives?).
5. The article says to code to the EJB spec, not the app server, then goes on to list all the WebSphere features to enable. Is this a J2EE article or an IBM sales pitch? Ever heard of JBoss?
6. Use entity beans? This is where the article really lost any credibility. No, use the DAO pattern, and implement your DAOs using either direct JDBC or better still, an ORM tool like Hibernate. How can a best practices doc not mention this amazing product?
7. Use a logging framework like JDK 1.4 logging or Log4J? At runtime, sure. But please only use the Jakarta Commons Logging API within your code; this isolates your app from the actual logging implementation used in deployment.
8. Always clean up after yourself. Good advice, but better still is to use a framework that cleans up automatically. For example, if using JDBC, use Spring's JdbcOperations API to avoid any possibility of Connections, PreparedStatements, or ResultSets being left open by developers.
The pracice
Use stateless session beans instead of stateful session beans,
seems to be very naive. The J2EE patterns (Sun's Pub.)book
explains when and when not to use satefull session. I have not found any blanket justification for always using the stateless session beans.
I have not read the article, only the synopsis here. However, it occurs to me that the author may be not-so-secretly conspiring to keep developers from moving off Websphere.
Andrew, You kinda missed the point of the best practices article.
Also, you emphasised only a part of what they wrote and deliberately(?) missed the other part.
1. It implies that the only viable view technologies are JSP and XML/XSL. What about the excellent Velocity and FreeMarker templating engines? They are great for preventing developers from embedding business and controller logic in views.
Of course you're right, there are more view technologies, but as you say, templating engines you refer to do the work for you, embedding business and controller logic in views. So, using MVC is a very good idea, apart from the technology you use, isn't it?
2. Use Struts? Maybe five years ago before Spring MVC came out. Spring's offering is so much more testable, extensible, and decoupled. I don't know much about Spring Web Flow, but it may be even better still, certainly for web workflows of any complexity.
You omitted the other part of the sentence ('Use Struts, Java Server Faces ...').
I personally think that JSF is a very promising technology. (with great frameworks like Seam gaining momentum).
3. Use Cactus to test J2EE components? I admit to doing this myself before seeing the light on POJOs. Write your application using POJOs, and you can test them using plain old JUnit without complicated and slow in-container testing. Then wrap them in a trivial EJB if you need to (although with a DI container like Spring you usually don't even need that).
Once again, the whole sentence used in the article says you should use Cactus OR JUnit. You can't just cut some words out ;-)
4. Use J2EE security for authorisation? For most real-world apps, where access is per-entity based on complicated criteria, it's useless (ref: www.owasp.org/index.php/Access_Control_In_Your_...). JAAS is only good for authentication. So overall, you're better off using a pluggable security framework like Acegi for both these aspects of security (are there any real alternatives?).
Of course, Acegi is often very good solution for security. But, I don't know where you tried to use Jaas in commercial application. I did, and I can say, that there are many possibilites there. You can write your own security realms and that allows you to use as sophisticated criteria as you want. I can't say it's easy, though.
5. The article says to code to the EJB spec, not the app server, then goes on to list all the WebSphere features to enable. Is this a J2EE article or an IBM sales pitch? Ever heard of JBoss?
Point taken, sounds like a sales pitch ;-).
6. Use entity beans? This is where the article really lost any credibility. No, use the DAO pattern, and implement your DAOs using either direct JDBC or better still, an ORM tool like Hibernate. How can a best practices doc not mention this amazing product?
Have you used the Java Persistance API and it's entity beans recently? It's very similar to Hibernate. In fact, you can use Hibernate as the JPA implementation provider. I also think that is has a big advantage over hibernate - it's standarised in the JEE 5 specification.
7. Use a logging framework like JDK 1.4 logging or Log4J? At runtime, sure. But please only use the Jakarta Commons Logging API within your code; this isolates your app from the actual logging implementation used in deployment.
Maybe I've some old habits, but I always used log4j for logging... Well, I'll try the commons logging :-).
8. Always clean up after yourself. Good advice, but better still is to use a framework that cleans up automatically. For example, if using JDBC, use Spring's JdbcOperations API to avoid any possibility of Connections, PreparedStatements, or ResultSets being left open by developers.
Well, they didn't write anywhere that you can't use anything to clean up for you ;-)
Regargs,
Luke
John Hugg discusses high volume transaction processing applications with high and low frequency profiles, and how VoltDB can be used for that purpose.
Kevlin Henney examines code samples to see what can be learned from them starting from the premise that one won’t write great code unless he knows how to read it.
Jason Ayers share the observations he made watching a team of developers collaborating in real time on the same code base, pushing XP, pair programming and continuous integration to their extremes.
Michael Snoyman presents Yesod, a web framework written in Haskell and containing a web server, templating, ORM, libraries (templating, gravatar, etc.).
Richard Kreuter and Kyle Banker on how to avoid classical RDBMS transactional systems by using compensation mechanisms, transactional messaging or transactional procedures.
Attila Szegedi talks about performance tuning Java and Scala programs at Twitter: how to approach GC problems, the importance of asynchronous I/O, when to use MySQL/Cassandra/Redis, and much more.
One category of risk that project teams need to ensure they address is business value failure – delivering a product that fails to provide value for the business investor.
InfoQ spoke to the authors of Software Systems Architecture on a couple of new topics, the System Context viewpoint and Agile, which have been added to the second edition.
4 comments
Watch Thread Reply