Claudiu Dumitrescu

Greenhorn
+ Follow
since Sep 09, 2008
Merit badge: grant badges
For More
Cows and Likes
Cows
Total received
0
In last 30 days
0
Total given
0
Likes
Total received
0
Received in last 30 days
0
Total given
0
Given in last 30 days
0
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Claudiu Dumitrescu

Hi,

I need to communicate with an external system using JMS.
Now, since my intention is to use EJB3 and have all the JMS objects (the ConnectionFactory and the Queue) injected via annotations by the EJB container, what level of details should I give to the sequence diagram?

I was thinking just to have a method in my EJB from which to create a self-call to depict the QueueSender creation and then just to call the remote Queue using a send method.
Is this acceptable or it doesn't make any sense? Should I depict QueueSender actual creation by adding the objects to the diagram?

To explain me better I add the diagram removing any other specific information to my case

Any help will be great as I'm stuck for two days on this not being able to decide which way to go.

Claudiu
I also will go for splitting by layer (behind Jeanne confirmation that would be OK).

Trying to split by feature will end, in my case, to an equivalent mess as I'll end up using a large amount of the business layer classes for each of the features. Another unwanted consequence will be duplication of classes on each diagram.

Then should I:

a) add a note stating the design patterns I used?

b) drop the all stereotypes that describes the various roles of the classes?

Thank you
Hi,

I'm thinking to mark the entities from the sequence diagram with various stereotypes like: <<Command>> , <<SessionFacade>> to describe the various patterns used.

What do you think about it?

Thank you.
Hi,

I am requested to offer a solution with high availability: but to what extent should I go with my deployment configuration?
I already have two reverse proxy and load balancing servers in an active-passive cluster, an active-active cluster of 4 application servers deployed on two nodes each on its own physical machine. Now, I will need a messaging server and a database. Should I go and cluster those two also? Am I going to far?

What deployment topology did you use for your assignment?

Thanks,
claudiu
Hi,

I'm considering having an HTTP server as a reverse proxy between the clients and the cluster of application servers.
This will require off-course a DMZ zone in witch this reverse proxy will sit.

My question is: how do I depict this DMZ on the deployment diagram?

I was thinking to add an artifact named <<firewall>>, on the the physical machine of the reverse proxy. But then where do I put the second firewall?

Or perhaps what I need are two more physical machines to install the firewall software on them? In this case I will finish having a deployment topology quite stuffed with servers:
2 for firewalls, 1 for reverse proxy, 2 for clustered application servers and perhaps one more for the Message Broker.

Is the price of the solution an issue I need to consider?
Have you been able to find the PetStore application? Because I wasn't. What it would interest me is the architecture documents. I found this for the Adventurer application (http://java.sun.com/blueprints/code/adventure/1.0/docs/architecture.html) but this is not using the EJB. Could someone point me to the right location for PetStore?
Hi,

It is my intention to use Front Controller together with Application Controller and to show the classes in the Class Diagram.
Since I will need to show the Commands objects involved the diagram it is going to grow behind one page.

It is to much if I split the Class Diagram in two: Controller Classes and Business Classes?

And one other question: it is safe to assume the session beans can be injected in the Commands objects or should I use a Service Locator also (given that I can use JEE5)?
Do you work on 'Retire Early' assignment? I am too.

I don't think we are supposed to change the business domain model.
Hi,

I have a scenario like this: an object will become available to SuD after some number of days after its creation. Taking in consideration that any loss of object means loss of money for the user, which of these two approaches do you think is more appropriate:

1) use a JMS Queue to be read every some period, to select the messages which are due to become available to the system and physically create them in database.

2) Or store them in the database right away after creation and use a scheduler to make them available by creating the necessary DB rows when the time is due

Second solution I think will pollute my DB's tables in some way since the domain model, as it was given, does not seems support this. And also will force me to a more complex management of domain objects even for simple scenarios.

On the other hand, the first solution rise the question of how reliable I should think the Message Broker System is. If this fails, I will loose all the objects which is unacceptable.
I also thought of using a temporary queue for my assignment, to model a request-response type of communication. But since it's going to be quite chatty I think the performance will be gravelly affected due to to creation of all these temporary queues. I decided to use a permanent queue with correlated ids and go for the message selector.
Yes, perhaps it better to keep them together. As a bonus I'll simplify the security issue as well.

Did anyone felt like using a JMS standalone server as part of deployment plan? Are any obvious detriments?

Hi,

For the part 2 I'm contemplating the possibility of a clustered ear deployment in order to assure a high availability. So it will be something like one web server and one application server with two (or possibly more) partitions.

- I think in this case I'll be forced to introduce the Service Locator at the web server level. What do you think?

- Does anyone know how should I go about the JMS Queues? How should I address the replication for these? It is a service usually assured by the application servers?

Thanks,
Claudiu
Hi,

I have one doubt and I hope you can help me: having a composition association, lets say between an Order and its Payment, does this imply that I cannot have an Order without an associated Payment? I know this holds true for the reverse but my opinion is that I should be able to have an Order without a Payment. I'm not sure though. Do I need to change the association style?

Thank you,
Claudiu
Hi there,

It seems we have the same assignment.

Anyway, I did a little research on the field, but this term does not appear to be a standard term in the business. My conclusion is that it could represent the assets an user can have. Actually later on is never referred by any of the use-case which speaks only of Assets. Therefore my conclusion. Hope it is right.

Good luck.
Claudiu