Sept. 7th, 1999: SEC Study Group: CORBA -- Part I -- Overview

• Moderator: liuj
• Scribe: liuj

• started: 2PM
  
• ended: 3:15PM

• Study materials
• Notes
• Related issues
• Concluding notes

Study materials

• The Common Object Request Broker: Architecture and Specification
  We will focus on Chapter 1, and C hapter 2.
  If you have more time, then you may also read Chapter 7, and Chapter 8.
  If you have even more time, you may read Chapter 3, Chapter 4, Chapter 5, and Chapter 6. And that concludes the CORBA core.

• S. Vinoski (vinoski@iona.com) CORBA: Integrating Diverse Applications Within Distributed Heterogeneous Environments

  Please, bring the materials to attend the study group.

Notes

• Study the OMA object model.
• Study the basic architecture of CORBA.
• Question: what is the basic step of CORBA programming?

  We will look at this in detail on this Friday. The basic steps are:
  

  1. Write the IDL definition of the servant object.
  2. On the server side, compile the IDL with a IDL compiler, which generate the "skeleton".
  3. Write the servant implementation, which implement the interface defined in the IDL.
  4. Start the servant, and register it to the ORB
  5. On the client side, compile the IDL with a IDL compiler, which generate the "stub".
  6. Write the client program, which treats the stub as the real servant.
  7. The client get an object reference from a string or a name server.

• Question: what is the size of a typical ORB?

  Leave to this Friday.

• Question: does CORBA provide real-time specifications, like QoS, priority, and scheduling?

  There are Real-time CORBA and Minimum CORBA SIG's in the CORBA community. People are getting serious about the real-time embedded applications using CORBA. The status of these communities needs to be checked.

• Question: What are the known-uses of CORBA?

  Need to be surveyed. Leave to Friday.

  Related issues

  1. People need to do their home work. We choose the study group format in order to get more people involved.

  Concluding notes

  People agree that CORBA focuses more on architectural flexibility than on performance issues. It is not obvious that it is suitable for real-time embedded systems. We hope to have a better understanding of this issue after the study groups on CORBA and TAO.

  We will continue on this Friday. The focus will be the problems raised today.

  ---

  comments to:
  liuj@eecs.berkeley.edu