Rodrigo Cunha

Originally posted by Deepak Bala:
I remember having to go through the same set of questions

Here is what I considered to come to a conclusion

* If we maintain sessions in the HTTP session scope, how would a rich client maintain session

* What will be the impact on scalability and performance if stateful session beans are used.

I had to ponder on those two questions for a while before coming to a conclusion. Good luck with your design

Thank you so much for such valuable reply

best regards

Originally posted by dhruboj kayal:
if you need colocated architecture do you think that EJB is required?

Unfortunately for reasons I can�t talk about.. yes.

Actually what is killing me here is that I need to keep track of user session, and doing so w/ Stateful session beans in a multi-tier approach may compromise performance (w/ lookups), that we could prevent with a co-located approach.

Now you might be thinking why not multi-tier and having HttpSession clustering for the web tier (since many vendors already provide it)? I strongly believe the test reviewers would be happier to see a solution with stateful session bean (which is blueprints recommendation when using EJBs) rathen then a solution with HttpSession clustering provided by a third party. But again, it�s just a guess.

Any comments?

Regards,

Originally posted by wang andrew:
If there is no mention of non web clients, then go http session.
Since performance is paramount requirement, why don't you choose web ejb co-located architecture?

makes sense, and it�s surely something to think about.. the only thing is that I�ve already proposed a DMZ which is where I�d expose the web servers and have protected the application servers inside the internal network (behind the 2nd firewall). I�d probably have to change that..

I�ll study the advantages and drawbacks of this approach and see how feasible it would be. will let you know.

Thank you so much for the reply.. really appreciate it.

best regards,

Hi,

My application is a web application which I chose to use JSF. Now, being that performance is a paramount requirement, would it be better to use a JSF managed bean with session scope or Stateful session beans to maintain conversational state with client?

I know that maintaining session state in stateful session beans is a BluePrints best practice, but aren't they too heavyweight when compared to HttpSession? Also, I've separated my web and EJB tiers to run on different servers and each server itself is clustered.. I know some web servers do provide HttpSession clustering, so I mentioned I'd be using one that could guarantee that.

Important notice: there's no mention of other non web applications (i.e. swing app) accessing the business tier of my application. This is very misleading when taking such decision.

Any comment will be greatly appreciated.

Best regards,

please check.. https://coderanch.com/t/156750/java-Architect-SCEA/certification/Exam-tips-feedback-Sun-epractice

Originally posted by Nikhil Jain:
Thats all the patterns of the book. So do we need to go through all of them?

yes! you have to know ALL them upside-down.. don�t save any effort.. this is not an easy test.

regards

good point Cameron..

yes, you shouldn�t forget to mention FacesServlet as your front controller. Now, I also had the Faces Lifecycle as my application controller, is this a good decision?

I noticed on the following article http://websphere.sys-con.com/read/46516.htm that JSF uses the Page Controller design pattern, any comments on that?

regards,

I'm actually modeling the Managed Beans as POJOs, but haven't submitted the project yet. So I'm not 100% sure this is the right approach.
Maybe someone who already did part 2 could reply this post..

Agree with Theodore.

Anurag, I went over the same dilemma a couple of months ago and I decided on taking the new SCEA 5. There's no point on taking the old one since it's being discontinued anyways. Also, they're not selling vouchers for that anymore and you'll have until december to finish all 3 parts.

But in order to prepare for it you'll have to work a lot harder to cover new topics that are not covered on the old exam. Also, the test is very different from the first one, so don't expect old exam simulators to help much.

Anyways, try focusing on some topics on this forum.. some of them will really help guiding you (specially those that post exam results).
Also, check my feedback topic on the new exam and on epractize labs exam simulator https://coderanch.com/t/156750/java-Architect-SCEA/certification/Exam-tips-feedback-Sun-epractice

One of the books I'd strongly suggest (besides those I mention on my feedback topic) is Practical Software Estimation: (M. A. Parthasarathy). This reference is mentioned all around the sun's epractice exam simulator.

Good luck, and keep in mind the test is a lot harder than the old version, so do not save efforts to prepare for it.

regards,

Originally posted by Suresh Panneer Selvam:
As I understand from you guys, SCEA 5 part II exam is assignment. Can you please tell some more details about how it is this exam? Will I download problem from Sun site, solve and then upload back into sun it or something else?

Thanks
Suresh P.
ITIL V3,SCJP, SCBCD, SCWCD, SCEA 5 Part I.

Exactly! You should contact Sun at your country so they'll guide you thru the process. You'll probably have to send them an email requesting your assignment, then they'll give you a login and password for http://www.certmanager.net/sun_assignment/ and you'll be able to download your assignment from there.
The file you download will contain every instruction you need to know to complete the assignment, after you're done, you'll have to submit it at the same website.

hope this helps,
regards,

Does your voucher come with retake option? If so, you need to check with Prometric, I believe there's a period of time you need to wait before taking it again. I'd suggest you work on what you feel your weeknesses are before taking it. And try some simulators as well.

best wishes..

First of all, I'd like to say I honestly admire your courage on having posted your unsuccessful experience with the exam here. I strongly believe there were many sharing the same experience that wouldn't have done the same.

Now, you shouldn't feel so bad about it.. the exam is really tough!! You don't actually see high scores, only people barely passing.

About the EPractize labs simulator, don't feel bad you haven't acquired it. I believe it wouldn't help you much... please have a look at my feedback post - https://coderanch.com/t/156750/java-Architect-SCEA/certification/Exam-tips-feedback-Sun-epractice

I felt exactly the same way about timing.. despite what many people have posted here that timing wasn't an issue, for me and for many people I know... it really was!

Another thing that didn't help much was the lack of material to study for the exam. That is really misleading.. what I did was trying to focus on the references from the Sun's epractice and what is suggested on sun's exam site.

Things like that can't happen without you learning from it.. so don't give up, now that you know what the exam looks like.. you have a hell of advantage. Try focusing on what will really help you passing...

Wish you very good luck!
best regards,
Rodrigo Cunha

Originally posted by Daniel Pariser:


Hmm, good question. As said, the practice question had mostly only one valid answer and this was in many cases the only feasible answer. You could very easily guess the correct answer once you familiarized with the asking style. This was not the case in the real exam. In the real exam, most questions had more than one answer, which made this a lot harder. The 2 minutes per question is hardly enough time to read and fully understand the question.

That said and looking back I think the epractice was a good place to familiarize with the style of the questions in order to not be shocked in the real exam.

Thanks

Thank you for the reply, I took the exam and actually found the time an issue.

have you finished second part?

regards

Originally posted by Paren Parikh:
Hello Rodrigo,

When you said that questions were scenario based so out of the 64 questions how many were scenario based ? All ? or 50 %?

Any input is highly appreciated..

Regards
Paren

I'd say 85% - 95% were scenario based.
Be prepared to know the advantages and drawbacks of using certain technology on a specific scenario that they'll come up with. I'm not allowed to give you a specific scenario here.. but they'll test your ability to understand a scenario and apply a set of technologies to it. That's why I think it's kinda hard to point out a "read and you'll pass" book for this certification.

Good luck and hope this help.
Regards

Originally posted by Vijayagopalan Raveendran:
Congrats !

I took the exam on June 27 2008, scored 62%, below my expectation.
The exam was a real tough one in the my certification series with sun.

Was really tested in the areas which I thought I was very strong.
I completed my other SUN certifications in 1HR, this one took almost 2HR to complete.

Really one has to study well and understand the concepts very clearly
esp
Architectures & NFRs
Design Patterns
Securtiy
Web Services
Messaging

Having prepared with out much assistance, the score did not matter to me.

Getting ready for the Part 2

Thanks,

Thank you all for your replies.. and yes! this one is really tough.. I felt the same way Vijayagopalan did. it really put me to the test on subjects that I felt I was pretty solid on knowledge.