]> CORBA CORBA to Java Mapping Jini Proxies Simple Example Corba Server in Java Corba Client in Java Jini Service Jini Server and Client Building the Example Running the Example CORBA Implementations Room Booking Example CORBA Objects Multiple Objects Exceptions Interfaces for Single Thin Proxy RoomBookingBridge Implementation Other Classes Building the Example Running the Example Migrating a CORBA Client to Jini Jini Service as a CORBA Object Summary CORBA is also an infrastructure for distributed systems. It is targeted at a slightly different ``space'' to Jini, and has some minor and major differences. It allows specification of objects which can be distributed. The concentration is on distributed objects rather than on distributed services It is language-independent, using an Interface Definition Language (IDL) for specification of interfaces CORBA objects can be implemented in a number of languages, including C, C++, SmallTalk and Java Current versions of CORBA pass remote object references, rather than complete object instances. Each CORBA object lives within a server, and the object can only act within this server. This is more restricted than Jini, where an object can have instance data and class files sent to a remote location and execute there. This limitation in CORBA may change in future with ``pass by value'' parameters to methods IDL is a language that allows the programmer to specify the interfaces of a distributed object system. The syntax is similar to C++, but does not include any implementation-level constructs. So it allows definitions of data types (such as structures and unions), constants, enumerated types, exceptions and interfaces. Within interfaces it allows declaration of attributes and operations (methods). The complete IDL specification can be found on the Object Management Group (OMG) site http://www.omg.org. The book ``Java Programming with CORBA'' by Andreas Vogel and Keith Duddy (http://www.wiley.com/compbooks/vogel) contains an example of a room booking service specified in CORBA IDL and implemented in Java. This defines interfaces for Meeting, a factory to produce them, MeetingFactory and a Room. A room may have a number of meetings in slots (hourly throughout the day), and there are support constants, enumerations and typedefs to support this. In addition, exceptions may be thrown under various error conditions. The IDL that follows differs slightly from that given in the book, in that definitions of some data types that occur within interfaces have been ``lifted'' to a more global level, since the mapping from IDL to Java has changed slightly for elements nested within interfaces since that book was written. CORBA has bindings to a number of languages. The most recent addition is Java, and this binding is still under active development (that is, the core is basically settled, but some parts are still changing). This binding must cover all elements of IDL. A horribly brief summary is:

Module

A module is translated to a java package. All elements within the module becomes classes or interfaces within the package

Basic types

Most of the basic types map in a straightforward manner: a CORBA int becomes a Java int, a CORBA string becomes a Java java.lang.String, and so on. Some are a little tricky, such as the unsigned types which have no Java equivalent.

Constant

Constants within a CORBA IDL interface are mapped to constants within the corresponding Java interface. Constants which are ``global'' have no direct equivalent in Java, and so are mapped to Java interfaces with a single field which is the value.

Enum

Enumerated types have no direct Java equivalent, and so are mapped into a Java interface with the enumeration as a set of integer constants.

Struct

A CORBA IDL structure is implemented as a Java class with instance variables for all fields.

Interface

A CORBA IDL interface translates into a Java interface.

Exception

A CORBA IDL exception maps to a final Java class.

This mapping does not conform to naming conventions such as those established for Java Beans. For example, the IDL declaration readonly string purpose becomes the Java accessor method String purpose() rather than String getPurpose(). Where Java code is generated, the generated names will be used. But in ``ordinary'' methods, I shall use the more accepted naming forms. A Jini service exports a proxy object that acts within the client, on behalf of the service. On the service provider side there may be service backend objects, completing the service implementation. The proxy may be ``fat'' or ``thin'', depending on circumstances. In the chapter on MindStorms, the proxy has to be thin, since all it does is pass on requests to the service backend, which is linked to the hardware device. The service cannot move since it has to talk to a particular serial port, so the proxy is thin. (The proxy may have an extensive user interface, but the Jini community seems to feel that any user interface should be in Entry objects rather than in the proxy itself.) Proxy objects created as RMI proxies are similarly thin, just passing on method calls to the service backend, implemented as remote objects. CORBA services can be delivered to any accessible client. Each service is limited to the server on which it is running, so they are essentially immobile. They can be found by a variety of methods, such as a CORBA naming or trading service. These search methods can be run by any client, anywhere. A search will return a reference to a remote object, which is essentially a thin proxy to the CORBA service. Similarly, if a CORBA method call creates and returns an object, then it will return a remote reference to that object which will continue to exist on the server where it was created. (The new CORBA standards will allow objects to be returned by value. This is not yet commonplace, and will probably be restricted to a few languages such as C++ and Java.) The simplest way to make a CORBA object available to a Jini federation is to build a Jini service that is at the same time a CORBA client. The service acts as a bridge between the two protocols. Really, this is just the same as MindStorms - anything that requires talking a different protocol (hardware or software) will require a bridge between the two. Most CORBA implementations use IIOP - that is, they are TCP based. The current Jini implementation is also TCP based. So there is a confluence of transport methods, which normally would not occur. A bridge would usually be fixed to a particular piece of hardware, but here it is not necessary due to this confluence. A Jini service has a lot of flexibility in implementation, and can choose to place logic in the proxy, in the backend, or any where else for that matter! The combination of Jini flexibility and IIOP allows a larger variety of implementation possibilities than can be done with fixed pieces of hardware such as MindStorms. For example The Jini proxy can invoke the CORBA naming service lookup to locate the CORBA service, and then makes calls directly on the CORBA service from the client. This is a ``fat'' proxy model in which the proxy contains all of the service implementation. There is no need for a service backend, and the service provider just exports the service object as proxy and then keeps the leases for the lookup services alive The Jini proxy can be an RMI stub, passing on all method calls to a backend service running as an RMI remote object in the service provider. This is a ``thin'' proxy with ``fat'' backend, where all service implementation is done on the backend. The backend uses the CORBA naming service lookup to find the CORBA service and then makes calls on this CORBA service from the backend The CORBA IDL tutorial at http://java.sun.com/products/jdk/1.2/docs/guide/idl/ gives a ``hello world'' example. This is defined by the IDL This can be compiled into Java using a compiler such as Sun's idltojava (or other Corba 2.2 compliant compiler). This results in a package corba.HelloApp containing a number of classes and interfaces. Hello is an interface which is used by a CORBA client (in Java). A server for the hello IDL can be written in any language with a CORBA binding, such as C++. Rather than get diverted into other languages, though, we shall stick to a Java implementation. However, this language choice is not forced on us by CORBA. The server must create an object that implements the Hello interface. This is done by creating a servant that inherits from the _HelloImplBase, and then registering it with the CORBA ORB (Object Request Broker - the CORBA ``backplane''). The server must also find a name server and register the name and the implementation. The servant implements the Hello interface. The server can just sleep to continue existence after registering the servant. A standalone client finds a proxy implementing the Hello interface by methods such as looking up a CORBA name server. The name server returns a org.omg.CORBA.Object which is cast to the interface type by the HelloHelper method narrow() (the Java cast method is not used). This proxy object can then be used to call methods back in the CORBA server In order to make the CORBA object accessible to the Jini world, it must be turned into a Jini service. At the same time it must remain in a CORBA server, so that it can be used by ordinary CORBA clients. So we do nothing to the CORBA server. Instead, we need to build a Jini service that will act as a CORBA client. This service will then be able to deliver the CORBA service to Jini clients. The Jini service can be implemented as a fat proxy delivered to a Jini client. The Jini service implementation is moved from Jini server to Jini client as the service object. Once in the client it is responsible for locating the CORBA service using the CORBA naming service, and then translates client calls on the Jini service directly into calls on the CORBA service. The processes that run in this, with their associated Jini and CORBA objects, are shown in figure .

The Java interface for this service is quite simple, and basically just copies the interface for the CORBA service, as The method sayHello() for the CORBA IDL returns a string. In the Java binding this becomes an ordinary Java String, and the Jini service can just use this type. The next example will show a more complex case where CORBA objects may get returned. Note that as a ``fat'' service, any implementation will get moved across to a Jini client and will run there. So the service only needs to implement Serializable, and its methods do not need to throw Remote exceptions since they will run locally in the client. The implementation of this Jini interface will basically act as a CORBA client. Its sayHello() method will contact the CORBA naming service, find a reference to the CORBA Hello object and call its sayHello() method. The Jini service can just return the string it gets from the CORBA service. The Jini server which exports the service doesn't contain anything new compared to the other service providers we have discussed. It creates a new JavaHelloImpl object and exports it using e.g. a JoinManager joinMgr = new JoinManager(new JavaHelloImpl(argv), ...) Similarly, the Jini client doesn't contain anything new, except to catch CORBA exceptions. After lookup discovery, the code is try { hello = (JavaHello) registrar.lookup(template); } catch(java.rmi.RemoteException e) { e.printStackTrace(); System.exit(2); } if (hello == null) { System.out.println("hello null"); return; } String msg; try { msg = hello.sayHello(); System.out.println(msg); } catch(Exception e) { // we may get a CORBA runtime error System.err.println(e.toString()); } The major additional step is to build the Java classes from the IDL specification. There are a number of CORBA IDL to Java compilers. One of these is the Sun compiler idltojava which is available from java.sun.com. This (or another compiler) needs to be run on the IDL file to produce the Java files in the package corba.HelloApp. These are standard Java files, and can be compiled using your normal Java compiler. They may need some CORBA files in the classpath if required by your vendor's implementation of CORBA. Files produced by idltojava do not need any extra classes. The Jini server, service and client are also normal Java files, and can be compiled like earlier Jini files, with the classpath set to include the Jini libraries. There are a large number of elements and processes that must be set running to get this running satisfactorily A CORBA name server must be set running. In the JDK 1.2 distribution is a server, tnameserv. By default this runs on TCP port 900. Under Unix access to this port is restricted to system supervisors. It can be set running on this port by a supervisor, or started during boot time. An ordinary user will need to use the option -ORBInitialPort port to run it on a port above 1024. For example, tnameserv -ORBInitialPort 1055 All CORBA services and clients should also use this port number The Java version of the CORBA service can then be started by java corba.CorbaHelloServer -ORBInitialPort 1055 Typical Jini support services will need to be running, such as a Jini lookup service, the RMI daemon rmid, and HTTP servers to move class definitions around The Jini service can be started by java corba.JavaHelloServer -ORBInitialPort 1055 Finally, the Jini client can be run by java client.TestCorbaHello -ORBInitialPort 1055 There are interesting issues about what is needed in Java to support CORBA. The example discussed above uses the CORBA API's that are part of the standard OMG binding of CORBA to Java. The packages rooted in org.omg are in major distributions of JDK 1.2, such as the Sun SDK. The example should compile okay with most Java 1.2 compilers with these OMG classes. Sun's JDK 1.2 runtime includes a CORBA ORB, and the example will run as is using this ORB. However, there are many implementations of CORBA ORB's, and they do not always behave in quite the same way. This can affect compilation and runtime aspects. The actual CORBA ORB used is determined at runtime, based on properties. If a particular ORB is not specified, then it defaults to the Sun supplied ORB (using Sun's SDK). To use another ORB, such as the Orbacus ORB, the following needs to be inserted before the call to ORB.init(): java.util.Properties props = System.getProperties(); props.put("org.omg.CORBA.ORBClass", "com.ooc.CORBA.ORB"); props.put("org.omg.CORBA.ORBSingletonClass", "com.ooc.CORBA.ORBSingleton"); System.setProperties(props); Similar code is required for the ORBS from IONA and other vendors. This could affect the runtime behaviour of the client: if the proxy expects to use a particular ORB apart from the default, then the class files for that ORB must be available to the client or be downloadable across the network. Alternatively, the proxy could be written to use the default Sun ORB, and then may need to make inter-ORB calls between the Sun ORB and the actual ORB used by the CORBA service. Such issues take us beyond the scope of this chapter, though! The IDL for a room booking problem was briefly discussed in an earlier section. This has a few more complexities over the previous example. The problem is to have a set of rooms, and for each room a set of bookings that can be made on that room. The bookings may be made on the hour, from 9am till 4pm (this only covers the bookings for one day). Bookings once made may be cancelled. A room can be queried for the set of bookings it has: it returns an array of meetings, which are null if no booking has been made, or a non-null meeting which has details of the participants and the purpose of the meeting. The additional features of this example are Each room is implemented as a separate CORBA object. There is also a ``meeting factory'' which produces more objects. This is a system with multiple CORBA objects residing on many CORBA servers. There are several possibilities for dealing with multiple objects Some of the methods return CORBA objects. These may need to be exposed to clients. This is not a problem if the client is a CORBA client, but here we will have Jini clients Some of the methods throw user-defined exceptions, in addition to CORBA-defined exceptions. Both of these need to be handled appropriately CORBA defines a set of ``primitive'' types such as integers of various sizes, chars, etc. The language bindings specify the primitive types they turn into for each language. For example, the CORBA wide character wchar becomes a Java Unicode char. Things are different for non-primitive objects, though. The IDL for the room booking defines CORBA interfaces for Meeting, MeetingFactory and Room. The implementation of these can be in any suitable language, and need not be in Java. The Java binding will convert these into Java interfaces. A CORBA client written in Java will get objects that implement these interfaces, but these will essentially be just references to remote CORBA objects. Two things are certain about these references: CORBA interfaces generate Java interfaces such as Hello. These inherit from org.omg.CORBA.portable.IDLEntity, which implements Serializable. So the references can be moved around like Jini objects. But they lose their link to the CORBA ORB that created them, and may end up in a different ``namespace'' where the reference makes no sense. So CORBA references cannot be usefully moved around. The closest one can come at present is to convert them to ``string-ified'' form and move that around. This may change when CORBA pass-by-value objects become common. Note that the serialization method that gives a string representation of a CORBA object is not the same as the Java one: the CORBA method serializes the remote reference, the Java method serializes the object's instance data They do not subclass from UnicastRemoteObject or Activatable. The Java runtime will not use an RMI stub for them If a Jini client gets local references to these objects, and keeps them local, then it can use them via their Java interfaces. If they need to be moved around the network, then appropriate ``mobile'' classes will need to be defined and the information copied across to them from the local objects. For example, the CORBA Meeting interface generates the Java interface To make the information from a CORBA Meeting available as a mobile Jini object, we would need an interface The first two methods allow information about a meeting to be accessible to applications that do not want to contact the CORBA service, and the third allows a CORBA object reference to be reconstructed within a new ORB. A suitable implementation is The implementation of the room booking problem in the Vogel and Duddy book runs each room as a separate CORBA object, each with its own server. A meeting factory creates meeting objects which are kept within the factory server, and passed around by reference. So for a distributed application with ten rooms, there will be eleven CORBA servers running. There are several possible ways of bringing this set of objects into the Jini world so that they are accessible to a Jini client, including A Jini server may exist for each CORBA server Each Jini server may export fat proxies, which build CORBA references in the same Jini client Each Jini server may export a thin proxy, with a CORBA reference held in each of these servers A single Jini server may be built for the federation of all the CORBA objects The single Jini server exports a fat proxy, which builds CORBA references in the Jini client The single Jini server exports a thin proxy, with all CORBA references within this single server The first of these essentially isolates each CORBA service into its own Jini service. This may be appropriate in an open-ended system where they may be a large set of CORBA services, only some of which are needed by any application. The second possibility deals with the case where services come logically grouped together, such that one cannot exist without the other, even though they may be distributed geographically. Intermediate schemes exist, where some CORBA services have their own Jini service, while others are grouped into a single Jini service. For example, rooms may be grouped into buildings and cannot exist without these buildings, whereas a client may only want to know about a subset of buildings, say those in New York. We can have one Jini server for each of the CORBA servers. The Jini servers can be running on the same machines as the CORBA ones, but there is no necessity from either Jini or the CORBA to do so. On the other hand, if a client is running as an applet, then applet security restrictions may force all the Jini servers to run on a single machine, the same one as an applet's HTTP server. The Jini proxy objects exported by each Jini server may be fat ones, which connect directly to the CORBA server. Thus each proxy becomes a CORBA client, as was dealt with in the ``hello world'' example. Within the Jini client, we do not just have one proxy, but many proxies. Because they are all running within the same address space, they can share CORBA references - there is no need to package a CORBA reference as a portable Jini object. In addition, the Jini client can just use all of these CORBA references directly, as instance objects of interfaces. Diagramatically, this appears in figure .

The CORBA servers are all accessed from within the Jini client. This may rule out this arrangement if the client is an applet, and the servers are on different machines. The proxies exported can be thin, such as RMI stubs. In this case each Jini server is acting as a CORBA client. This is given in figure .

If all the Jini servers are collocated on the same machine, then this becomes a possible architecture suitable for applets. The downside of this approach is that all the CORBA references are within different JVMs. In order to move the reference for a meeting from the Jini meeting factory to one of the Jini rooms, it may be necessary to wrap it in a portable Jini object, as discussed above. The Jini client will also need to get information about the CORBA objects, which can be gained from these portable Jini objects. An alternative to Jini servers for each CORBA server is to have a single Jini server into the CORBA federation. This can be a feasible alternative when the set of CORBA objects form a complete system or module, and it makes sense to treat them as a unit. There are then the choices again of where to locate the CORBA references, either in the Jini server or in a proxy. Placing them in a fat proxy is shown in figure

Placing all the CORBA references on the server side of a Jini service means that a Jini client only needs to make one network connection to the service. This is shown in figure . it is probably the best option from a security viewpoint of a Jini client.

CORBA methods can throw exceptions of two types: system exceptions and user exceptions. System exceptions subclass from RuntimeException, and so are unchecked. These do not need to have explicit try...catch clauses around them. If an exception is thrown, it will be caught by the Java runtime, and generally halt the process with an error message. This would result in a CORBA client dying, which would generally be undesirable. Many of these system exceptions will be caused by the distributed nature of CORBA objects, and probably should be caught explicitly. If they cannot be handled directly, then to bring them into line with the Jini world, they can be wrapped as ``nested exceptions'' within a Remote exception and thrown again. User exceptions are declared in the IDL for the CORBA interfaces and methods. These exceptions are checked, and need to be explicitly caught (or re-thrown) by Java methods. If a user exception is thrown, this will be because of some semantic error within one of the objects, and will be unrelated to any networking or remote issues. They should be treated as they are, without wrapping them in Remote exceptions. This and the following sections consider issues in building a single thin proxy for a federation of CORBA objects. The Vogel and Duddy book gives a CORBA client to interact with the CORBA federation, and this is used as the basis for the Jini services and clients. Using a thin proxy means that all CORBA-related calls will be placed in the service object, and made available to Jini clients only by means of portable Jini versions of the CORBA objects. These portable objects are defined by the interfaces and The bridge interface between the CORBA federation and the Jini clients has to provide methods for making changes to objects within the CORBA federation, and to obtain information from them. For the room booking system this requires the ability to book and cancel meetings within rooms, and also the ability to view the current state of the system. Viewing is accomplished by three methods: updating the current state, and getting the list of rooms and of bookings for a room. There is a slight legacy in this that comes from the original ``monoblock'' CORBA client. In there, because the GUI interface elements and the CORBA references were all in the one client, simple shareable structures such as arrays of rooms and arrays of meetings were used. Meetings and rooms could be indentified simply by their index into the appropriate array. In splitting the client apart into multiple (and remote) classes this is not really a good idea anymore as it assumes a commonality of implementation across objects which may not occur. It doesn't seem worthwhile being too fussy about that here, though. The room booking Jini bridge has to perform all CORBA activities, and to wrap these up as portable Jini objects. A major part of this is locating the CORBA services, which here are the meeting factory and all of the rooms. We do not want to get too involved in these issues here. The meeting factory can be found in essentially the same way as the hello server was earlier, by looking up its name. Finding the rooms is harder as these are not known in advance. Essentially, the equivalent of a ``directory'' has to be set up on the name server, which is known as a ``naming context''. Rooms are registered within this naming context by their servers, and the client gets this context and then does a search for its contents. The Jini component of this object is that it subclasses from UnicastRemoteObject, and implements a RemoteRoomBookingBridge which is a remote version of RoomBookingBridge. It is also worthwhile noting how CORBA exceptions are caught and wrapped in Remote exceptions. The Java classes and servers implementing the CORBA objects are mainly unchanged from the implementations given in the Vogel and Duddy book. They can continue to act as CORBA servers to the original clients. I replaced the ``easy naming'' naming service by a later one with the slightly more complex standard mechanism for creating contexts and placing names within this context. This can use the tnameserv CORBA naming server, for example. The Vogel and Duddy room booking client was messed around a little bit, but its essential structure remained unchanged. The GUI elements, for example, were not altered. All CORBA-related code was removed from the client and placed into the bridge classes. The Vogel and Duddy code samples can all be downloaded from a public Web site (http://www.wiley.com/compbooks/vogel), and come with no author attribution or copyright claim. The client is also quite lengthy since it has plenty of GUI inside. So I won't complete the code listing here. The code for all my classes, and the modified code of the Vogel and Duddy classes, is given in the subdirectory corba of the programs.zip file. The IDL interface RoomBooking.idl needs to be compiled to Java by a suitable ``IDL to Java'' compiler such as Sun's idltojava. This produces classes in the package corba.RoomBooking. These can all be compiled using the standard Java classes and any CORBA classes needed. The Jini server, service and client are also normal Java files, and can be compiled like earlier Jini files, with the classpath set to include the Jini libraries. There are a large number of elements and processes that must be set running to get this running satisfactorily A CORBA name server must be set running, as in the earlier problem. For example, tnameserv -ORBInitialPort 1055 All CORBA services and clients should also use this port number A CORBA server should be started for each room, with the first parameter being the ``name'' of the room, as in java corba.RoomBookingImpl.RoomServer "freds room" -ORBInitialPort 1055 A CORBA server should be started for the meeting factory java corba.RoomBookingImpl.MeetingFactoryServer -ORBInitialPort 1055 Typical Jini support services will need to be running, such as a lookup service, the RMI daemon rmid, and HTTP servers to move class definitions around The Jini service can be started by java corba.RoomBookingImpl.RoomBookingBridgeServer -ORBInitialPort 1055 Finally, the Jini client can be run by java corba.RoomBookingImpl.RoomBookingClientApplication -ORBInitialPort 1055 Both of the examples in this chapter started life as pure CORBA systems written by other authors, with CORBA objects delivered by servers to a CORBA client. The clients were both migrated to Jini clients of a Jini service acting as a front-end to CORBA objects, in a series of steps. For those in a similar situation, it may be worthwhile to spell out the steps I went through in doing this for the room booking problem. The original client was a single ``monoblock'' client, mixing GUI elements, CORBA calls and glue to hold it all together. The first step was to decide on the architectural constraint: one Jini service, or many. A single Jini service was chosen (for no other reason than it looked to offer more complexities). This implied that all CORBA related calls had to be collected into a single object, the RoomBookingBridgeImpl. At this stage, the RoomBookingBridge interface was not defined: that came after the implementation was completed (ok, I hang my head in shame, but I was trying to adapt existing code rather than starting from scratch). At this time, the client was still running as a pure CORBA client - no Jini mechanisms had been introduced. Once all the CORBA related code was isolated into one class, another architectural decision had to be made, as to whether this was to function as a fat or thin proxy. The decision to make it thin in this case was again based on interest rather than functional reasons. The GUI elements left behind in the client needed to access information from the CORBA objects. In the thin proxy model, this meant that portable Jini objects had to be built to carry information out of the CORBA world. This led to the interfaces such as JavaRoom and the implementations as JavaRoomImpl. The GUI code in the client had no need to directly modify fields in these objects, so they ended up as ``readonly'' versions of their CORBA sources. (If a fat proxy had been used, this step of creating portable Jini objects would not have been necessary.) The client was modified to use these portable Jini objects, and the RoomBookingBridgeImpl was changed to return these from its methods. Again, this was all still done within the CORBA world, and no Jini services were yet involved. This looked like a good time to define the RoomBookingBridge interface, when everything had settled down. Finally, the RoomBookingBridgeImpl was turned into a UnicastRemoteObject and placed into a Jini server. The client was changed to lookup a RoomBookingBridge service rather than create a RoomBookingBridgeImpl object. At the end of this, I had an implementation of a Jini service, with a thin RMI proxy. The CORBA objects and servers had not been changed at all. The original CORBA client had been split into two, with the Jini service implementing all of the CORBA lookups. These were exposed to the client through a set of ``facades'' that gave it the information it needed. The client was still responsible for all of the GUI aspects, and so was acting as a ``knowledgeable'' client. If needed, these GUI elements could be placed into Entry objects, and also exported as part of the service. We have looked at making CORBA objects into Jini services. Is it possible to go the other way, and make a Jini service appear as a CORBA object in a CORBA federation? Well, it should be. Just as there is a mapping from CORBA IDL to Java, there is also a mapping of a suitable subset of Java into CORBA IDL. So a Jini service interface can be written as a CORBA interface. A Jini client could then be written as the implementation of a CORBA serve to this IDL. At present, with a paucity of Jini services, it does not seem worthwhile yet to explore this in detail. This may change in the future, though. CORBA is a separate distributed system to Jini. However, it is quite straightforward to build bridges between the two systems, and there are a number of different possible architectures. This makes it possible for CORBA services to be used by Jini clients. ©right;