resubmit the deployment diagram - suggestions needed

[Lyju Edwinson]

Hi,
I have to resubmit the deployment diagram, this is my results below,

Questions 22 of 24
Class Diagram 39 of 40
Component Diagram 40 of 40
Deployment Diagram 14 of 24 ( Incomplete Deployment Diagram.)
Interaction Diagram 16 of 16
Technical Risks 14 of 16

Will you please assist me and give me some ideas like what is the expectation on Deployment Diagrams?
I followed Mark Cade's way of doing it.

Please give me some good suggestions.

Regards
Lyju

[Gurunath Dharmar]

Hi

what was the assignment and when did you submitted and when did you receive the result?

what was the information you have provided in deployment diagram? so that anyone can discuss on missing items.

[Gurukant Desai]

I am impressed by your score in sections other than deployment diagram!! Anyway, hard luck for deployment diagram, what was your assignment? I am also surprised by the reason stated "deployment diagram is incomplete"!!

[Lyju Edwinson]

Hi ,

Thanks for responding to my query.

I submitted the Assignment by 17th August 2011 and got results from pearsonvue.com.

Details below,
I have three layers in the deployment diagram with 12 boxes,

First layer shows the Web server Nodes ( Primary and secondary and COB). All are marked with Sterio types <<WebServer>>, <<WebServer Secondary>>, <<WebServer COB>>.

Second Layer Shows the Application Server Nodes. All are marked as Primary, secondary and COB as above. Both these layers are linked with Sterio Type <<RMI>>.

Third layer: Consists of DB server Node, Email Server and Other External Web Service End Point Nodes. All are marked with Corresponding Sterio Types. For example, the connection from Application Server to Web Service End point is shown as <<SOAP>>.

I haven't put things like appname.jar, .war, .ear etc in the nodes.

Please let me know if you need any other details.

Regards
Lyju

[Jari Timonen]

Do not copy Cades. It's incomplete. (I failed because of this.. used same style that in Cades)

Just use google image search: Deployment Diagrams in Google image search

[Lyju Edwinson]

Hi Thank you Timonen for the reply,

I think that is the mistake I made

I hope this one is a better approach, http://www.agilemodeling.com/artifacts/deploymentDiagram.htm

Regards
Lyju

[Giuseppe Dell'Abate]

I followed Mark Cade's approach for the Deployment Diagram and I am still waiting for the response.
Anyway in my Deployment I have added even the Firewall and the Load Balancer.
For each node I have create an hardware profile as a stereotype and I have explained: CPU, RAM, HDD, RAID, Network Adapter and a quick summary of the machine as a UML Note.
I have added another UML Note for each Tier where I have explained the main caracteristics, critics point and so on, not so much word but just to explain better the reason of my hardware choise.
I haven't added components and artifacts inside the nodes but I have explained in the UML Note the usage of the node.
My application receive request from external but it doesn't request external resource (web services) so I had less security problems.

Do you provide a Load Balancer and a Firewall?
Which type of algorith do you use in the Load Balancer ?
Do you explained the use of HTTP/HTTPS to access the Web Server ?
In your SuD you call an external system (web service) without pass through a firewall, DMZ ?

I don't know if my diagram is ok because my assignment is still in grading but this is what I have done.

Regards
Giuseppe

[Marius Cimpean]

Some additional info:
you should also provide a "study case", where you can defense your hardware architecture. For example, you should consider total number of users (concurrent) accessing the system, and total number of entities that are managed through your system. Yous should also provide some info about amount of memory used on runtime (session state / management). This way you could defense your decisions related to storage disk and RAM.

[Usman Ibrahim]

Have a look into this url. All the Best.

[Lyju Edwinson]

Hi,

Thank you all for the great assistance.
I am going to resubmit it on this Saturday.
I gone through the following articles, and feel a bit more confident.

http://www.uml-diagrams.org/deployment-diagrams-overview.html
http://www.ibm.com/developerworks/rational/library/769d.html
http://www.agilemodeling.com/artifacts/deploymentDiagram.htm

I hope I can clear this time......


Regards
Lyju

[Ganesh Gopalakrishnan]

My Few cents :

As suggested by Cade and Sheil in chapter 9 of their book, consider the below for deployment diagram.

[1] CPUs (number of cores, clock speed);
[2] RAM (quantity in GB);
[3] Network (minimum interface speed); and
[4] Storage (disk / SAN configuration).

I did some research and analysis and provided above details in my assignment.
I also followed cade and sheil approach for deployment diagram from a layout perspective, however did some(in fact lot of) research to do hardware sizing,
software(and their hardware configuration) to be deployed in different tiers etc

In fact, you can give thought on cloud based implementation too, for ex Amazon EC2,Rackspace etc.

Below link has PDF presentation by Humphery Sheil.
http://www.box.net/shared/i5eoqntdaf ( Look at slide 22 and 21&23(these slides need attention, may be in your case) ).
There could be many ways to do it, but getting aligned to evaluation criteria helps, is n't it?

For RAID, look at this : http://en.wikipedia.org/wiki/Standard_RAID_levels, you can get an idea and design it accordingly.

You might want to look at the below links as well.

1) http://stackoverflow.com/questions/5056657/how-to-convert-concurrent-users-into-hits-per-second
2) http://www.dba-oracle.com/concepts/database_administration.htm

I just gave some pointers, apparently, I wont be able to discuss beyond certain details...

All the Best!

[Lyju Edwinson]

Thank you for the details.
Let me go through that as well.

Regards
Lyju

[Lyju Edwinson]

Hi All,

I Resubmitted the assignment today.
Again Waiting .......

Thank you all for the details and great help.

Regards
Lyju

[Lyju Edwinson]

I can see my result from certview, I think I am passed


03-OCT-11 1Z0-867 Java Enterprise Edition 5 Enterprise Architect Certified Master Assignment for Resubmission PASS

18-AUG-11 1Z0-866 Java Enterprise Edition 5 Enterprise Architect Certified Master Essay Exam TAKEN

21-JUL-11 1Z0-865 Java Enterprise Edition 5 Enterprise Architect Certified Master Downloadable Project FAIL

22-JUN-11 1Z0-864 Java Enterprise Edition 5 Enterprise Architect Certified Master Exam (Step 1 of 3) PASS



Thanks a lot for the Help all........................

Regards
Lyju

[Usman Ibrahim]

Excellent. Congrats.

[Ganesh Gopalakrishnan]

Great!!! Good to Know....

[Lyju Edwinson]

Thank you Ganesh, those links really helped me.

Have a great day!
Cheers!

[Ganesh Gopalakrishnan]

Lyju! You are Welcome!

BTW, what is your assigment, just curious to know...

[Lyju Edwinson]

My Assignment was TeamDoctor.

Cheers
Lyju

[Viggo Navarsete]

Hi Lyju Edwinson,

I also have to resubmit my Assignment. Where did you send it, and how did you order it? Did you order the same assignment as the first time? or something else?

[Lyju Edwinson]

Hi Viggo,

I purchased the 1Z0_867 from http://www.pearsonvue.com/oracle/ and resubmitted the diagram again through the same site.

You don't need to resubmit a new Assignment. From your results you will come to know which areas has failed to meet the required mark.
Just focus on that part and resubmit the same modified diagrams for the same assignment.

Let me know if you have any more queries and All the very best.

Regards
Lyju

[Viggo Navarsete]

thanks for quick reply:) I could improve in a few areas:
Questions (19/24): More specific detail required for few questions. Which questions?
Class Diagram (23/40): Missing all public method names, Missing EJBs/explanation, and missing some interfaces
Component Diagram (30/40): Incorrect/missing UML notation for interface/implementation
Deployment Diagram (10/20): More detail required, Missing/confusing deployment of components/ear/war/ejb-jar,Firewall deployment missing details, Missing #(virtual) CPUs assigned per tier.
Interaction diagram (11/16): More detail required
Technical risks: More detail required for more than 2 risks.

So, I should probably deliver the whole package again, with all the diagrams, packaged as a JAR, right?

Any, not-to-detailed, input you could give?

[Lyju Edwinson]

Hi Viggo,
To pass each part you need to meet the minimum marks,

Comp D: 26/40
Clas D: 26/40
DD: 17/24
Intera: 0/16
Risk : 0/16
Questions (Essay): 17/24

I can see you passed the below ones,

Questions (Essay)
Component Diagram
Interaction Diagram

I think for risks it should be fine, because you need a total of 114 only to pass the exam.

So you only need to focus on the class diagram and deployment diagram here.

I would suggest you to drop an email to Oracle Support and confirm the same. ( I am pretty sure )

Let me know if you need more help.

Regards
Lyju

[Viggo Navarsete]

Hi Lyju,
I saw that you wrote
Intera: 0/16
Risk : 0/16

Was that intentionally? I suppose you need to score some points in those areas as well?

[Lyju Edwinson]

What that means is, there is no minimum barrier for the interaction diagrams and Risk sections.
But you can score Maximum there

Regards
Lyju

[gekko kamen]

Hi Lyju,

What is the process if you don't get the 17 points in the Essay Questions?

thank you.

[gekko kamen]

hi Lyju,

Other question I have, is if your class, component and sequence diagrams were based on the Cades book?
I made everything based on the book, but looks like they want to see a really more detailed design. I failed almost everything with the comment "Diagram is incomplete (all use cases not
covered). More details required". Does anyone have a recommendation to me?

thank you.

[Lyju Edwinson]

Hi kamen,

Honestly I don't know what to do if you failed in essay .
I guess there are some other friends also asked the same question in other topics here. Please do a search.

For the diagrams,

Cades book is a good reference. But what I did was, I googled for each diagrams and read small articles from IBM web site and javaworld.com. Wiki also was a good source.

Once I gone through these sites I got ideas on how to go with the diagrams. So basically Cades book helped me to understand what is expected in the exam.

I also didn't put helper classes and other java beans in the class diagram. I just mentioned in notes that I am not including those in class diagram to control size.

So My class diagram contains,

1) Main JSPs
2) The controller class. (Servlet)
3) BOs
4) SFacades
5) Entities.

Similarly I tried to include the required information only in all the other diagrams. For example I didn't put login sequences in the sequence diagram. I just put seq diagrams for all use cases and mentioned in notes that login sequence is not included to control size.

Basically whatever you follow, you need to make sure that your diagrams express your ideas and patterns. Thats all.

Let me know if you need anything more...

Also use a good tool to draw the diagrams. I used 'astah' tool and is excellent one.

Regards
Lyju

[gekko kamen]

Hi Lyju,

I really appreciate your response. Even though you had to resubmit, you made a great job. Congratulations for that

I have for now some more questions, and I hope you can help me:

1. Is it necessary to include alternate scenarios and things like exception handling in the sequence diagrams?
2. When they say "augment your diagram with text" they mean some text as notes in UML (I guess this is not enough to show the detail required), or some paragraphs in html text, or a description of each class, component or element in the diagram?
3. In the component diagram, did you include all the JPA entities?
4. Did you show somewhere the "Transfer Objects" or is enough to show the Entities related to the business and DAO components?

thank you.

[Lyju Edwinson]

Hi Kamen,

Thanks and Please find the reply inline,

1. Is it necessary to include alternate scenarios and things like exception handling in the sequence diagrams?
--> I think by alternate scenarios you mean to go to a different code of execution based on some conditions, it is better to include those details. I didn't include Exception details
But if you think you have a new use case to meet the correct requirement, then add a new sequence diagram for that.

2. When they say "augment your diagram with text" they mean some text as notes in UML (I guess this is not enough to show the detail required), or some paragraphs in html text, or a description of each class, component or element in the diagram?
--> No need to put comment for each class. But you need to put a comment if that really make a point there. For example if your Class diagram has a class Authenticator.java, then put a note saying what type of authentication used there. I hope you got what I meant.

3. In the component diagram, did you include all the JPA entities?
Not all the entities, but most of them. For example User was an entity with Doctor and Specialist as sub classes. I included User only in component diagrams.
4. Did you show somewhere the "Transfer Objects" or is enough to show the Entities related to the business and DAO components?
No I didn't, if you use the Business Objects or Session Facades then include that. Just mention as a separate note that you excluded transfer objects and other helper classes for brevity.

Let me know if this make sens to you ...

Regards
Lyju

[gekko kamen]

hi Lyju, great help, you are so kind.

some other questions:

1. In the Class diagram; I use a generic class to represent a group of classes of the domain model. I think this approach is ok, because the system will grow up, and I don't want to add new classes and make a new deployment when a new "Type" of "Something" is added.
Am I right? or is the intention of this Class diagram to represent exactly the same entities as in the "Domain Model" provided, but with attributes and navigation specified?

2. If I made a class diagram framework agnostic, as suggested in Cade & Sheil book, the sequence and component diagram should be also framework agnostic? would be better to specify the frameworks I will use in these diagrams?

3. In the deployment diagram I just made 14 point as you did at first. I guess that almost the same diagram with a deeper explanation on why I made that desicions would be enough. Do you think it is necessary to make a big change to the Cade's approach?

4.This is a question about the presentation of the work. I want to know if UML notes are enough to make all the comments. Did you need to add some paragraphs in html to achieve this?

thank you, your help has been so helpful.

[Thomas Felix]

I failed in Deployment diagram. Need to resubmit it.

I need to make jar with updated deployment diagram only?
Or with other old diagram and new deployment diagram?

With regards,
Thomas T

[ash Pandey]

Send the whole jar with all the updated components.

[Lyju Edwinson]

@Kamen,

Please see my comments inline

1. In the Class diagram; I use a generic class to represent a group of classes of the domain model. I think this approach is ok, because the system will grow up, and I don't want to add new classes and make a new deployment when a new "Type" of "Something" is added.
Am I right? or is the intention of this Class diagram to represent exactly the same entities as in the "Domain Model" provided, but with attributes and navigation specified?
You can verywell put the Generic Class or interface in the Class diagram to trepresent the group. I did the same but I made it sure that all public methods are declared in the diagram properly. Also i put the attributes that are very necessary. For example if you have a variable expecting an injected object, then declare it and mention that <<Injected Value>> some thing like that. If your class diagram is readable with proper notations then you get full marks.

2. If I made a class diagram framework agnostic, as suggested in Cade & Sheil book, the sequence and component diagram should be also framework agnostic? would be better to specify the frameworks I will use in these diagrams?
Yes I have not followed any frame work, I just mentioned in notes some of the favourable frame works that supports my application and why. Frame work is not necessary (depends on your app and requirement).

3. In the deployment diagram I just made 14 point as you did at first. I guess that almost the same diagram with a deeper explanation on why I made that desicions would be enough. Do you think it is necessary to make a big change to the Cade's approach?
Yes, you need to make a big changes in cadde's approach. Please look at this topic above for the discussions and suggestions by our friends. If you spend One day and try some DDs generally you can make a very good DD.

4.This is a question about the presentation of the work. I want to know if UML notes are enough to make all the comments. Did you need to add some paragraphs in html to achieve this?

UML notes are enough. Because the diagram should be self explanative. If you have something to write in a paragaraph about your diagram, that implies that your diagram is incomplete.

Cheers
Lyju

[gekko kamen]

thanks,

I've recieved a passing score, so now I'm a SCEA
To be honest, I followed most of your advices, but mainly I followed my criteria (several approaches are valid for the same problem) and I focused on support and explain my decisions.

Good luck.

[Lyju Edwinson]

Congratulations Kamen, Welcome to the SCEA club .

Cheers!

[Viggo Navarsete]

Gekko: Congratulations!! I'm resubmitting my assignment tonight, do you have any last words about the deployment diagram to share (without getting to specific!)? Can you tell how you visualized clusters? As many nodes or as one node with a description/stereotype cluster? Did you add much description to your deployment diagram?

And what about the deployment of the WAR/EAR to the container, were you specific about it?

[gekko kamen]

Viggo Navarsete wrote:Gekko: Congratulations!! I'm resubmitting my assignment tonight, do you have any last words about the deployment diagram to share (without getting to specific!)? Can you tell how you visualized clusters? As many nodes or as one node with a description/stereotype cluster? Did you add much description to your deployment diagram?

And what about the deployment of the WAR/EAR to the container, were you specific about it?

Hi Viggo,

I guess that if you show one or more nodes to represent a cluster is not really important, you should choose the option you feel is more clear and makes the diagram more readable.
I prefer to add a description for each node in my diagram and be explicit with the protocols for comunication with each other. I wrote the reasons for my decisions based on de NFR of the system, this is what I consider important.

God luck with your assignment