1. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 1 Distributed Systems Architectures l Architectural design for software that executes on more than one processor

2. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 2 Topics covered l Multi-processor architectures l Client-server architectures l Distributed object architectures l CORBA

3. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 3 Distributed systems l Virtually all large computer-based systems are now distributed systems l Information processing is distributed over several computers rather than confined to a single machine l Distributed software engineering is now very important

4. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 4 System types l Personal systems that are not distributed and that are designed to run on a personal computer or workstation. l Embedded systems that run on a single processor or on an integrated group of processors. l Distributed systems where the system software runs on a loosely integrated group of cooperating processors linked by a network.

5. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 5 Distributed system characteristics l Resource sharing l Openness Can be extended by adding new non-proprietary components l Concurrency l Scalability l Fault tolerance l Transparency User does not know that the system is distributed

6. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 6 Distributed system disadvantages l Complexity l Security l Manageability l Unpredictability

7. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 7 Issues in distributed system design l Resource Identification l Communications l Quality of Service l Software Architectures

8. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 8 Resource identification l A naming scheme is needed so that users can discover and refer to resources Example: URL (Uniform Resource Locator) »Problem: Objects cannot be relocated. »URN: Universal resource name l Identification by service type (“printer”, “file server”, “certificate authority”, “name server”)

9. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 9 Communications l TCP/IP networking has become the standard communications mechanism. However, some applications require new protocols: Wireless Multi-cast Audio/video

10. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 10 Notes l Revised User Manual will be accepted for extra credit Revised user manual is no longer a requirement for the System Requirements assignment

11. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 11 Quality of Service l QoS: Performance, reliability, availability l Depends on many factors: The allocation of processes to processors in the system The distribution of resources across the system, The network The system hardware The adaptability of the system.

12. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 12 Software Architectures l The software architecture describes how the application functionality is distributed over a number of logical components and how these components are distributed across processors. l Choosing the right architecture for an application is essential to achieve the desired quality of service.

13. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 13 Distributed systems architectures l Multiprocessor architectures Parallel computing across multiple processors. Efficient, trusted communication often over dedicated network. l Client-server architectures Distributed services which are called on by clients. Servers that provide services are treated differently from clients that use services l Distributed object architectures No distinction between clients and servers. Any object on the system may provide and use services from other objects

14. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 14 Multiprocessor architectures l Simplest distributed system model l System composed of multiple processes which execute on different processors to achieve high performance l Communication by trusted, high-speed network l Architectural model of many large real-time and scientific computing systems l Distribution of processes to processors may be fixed or may be under the control of a dispatcher l Typically “closed” systems made of custom components.

15. A multiprocessor traffic control system

16. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 16 Distributed systems architectures (2) l Multiprocessor architectures l Client-server architectures l Distributed object architectures

17. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 17 Client-server architectures l The application is modeled as a set of services that are provided by servers and a set of clients that use these services l Clients know of servers but servers need not know of clients l Clients and servers are logical processes l The mapping of processors to processes is not necessarily 1 : 1

18. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 18 A client-server system

19. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 19 Computers in a C/S network

20. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 20 Layered application architecture l Presentation layer Concerned with presenting the results of a computation to system users and with collecting user inputs l Application processing layer Concerned with providing application specific functionality e.g., in a banking system, banking functions such as open account, close account, etc. l Data management layer Concerned with managing the system databases

21. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 21 Application layers

22. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 22 Thin and fat clients l Thin-client model In a thin-client model, all of the application processing and data management is carried out on the server. The client is simply responsible for running the presentation software. l Fat-client model In this model, the server is only responsible for data management. The software on the client implements the application logic and the interactions with the system user.

23. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 23 Thin and fat clients

24. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 24 Thin client model l Used when legacy systems are migrated to client server architectures. The legacy system acts as a server in its own right with a graphical interface implemented on a client l A major disadvantage is that it places a heavy processing load on both the server and the network

25. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 25 Fat client model l More processing is delegated to the client as the application processing is locally executed l Most suitable for new C/S systems where the capabilities of the client system are known in advance l More complex than a thin client model especially for management. New versions of the application have to be installed on all clients

26. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 26 Fat client example: A client-server ATM system

27. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 27 Three-tier architectures l In a three-tier architecture, each of the application architecture layers may execute on a separate processor l Allows for better performance than a thin- client approach and is simpler to manage than a fat-client approach l A more scalable architecture - as demands increase, extra servers can be added

28. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 28 A 3-tier C/S architecture l Presentation and Data Management can use generic tools (browser and DBMS) l Application-specific logic only in one location

29. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 29 An internet banking system

30. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 30 Use of C/S architectures

31. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 31 Distributed systems architectures (3) l Multiprocessor architectures l Client-server architectures l Distributed object architectures

32. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 32 Distributed object architectures l No distinction in a distributed object architecture between clients and servers l Each distributable entity is an object that provides services to other objects and receives services from other objects l Object communication is through a middleware system called an object request broker (software bus) l However, more complex to design than C/S systems

33. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 33 Middleware l Software that manages and supports the different components of a distributed system. In essence, it sits in the middle of the system l Middleware is usually off-the-shelf rather than specially-written software l Examples Transaction processing monitors Data convertors Communication controllers

34. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 34 Distributed object architecture

35. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 35 Advantages of distributed object architecture l It allows the system designer to delay decisions on where and how services should be provided l It is a very open system architecture that allows new resources to be added as required l The system is flexible and scaleable l It is possible to reconfigure the system dynamically with objects migrating across the network as required

36. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 36 Uses of distributed object architecture l As a logical model that allows you to structure and organize the system. In this case, you think about how to provide application functionality solely in terms of services and combinations of services l As a flexible approach to the implementation of client-server systems. The logical model of the system is a client-server model, but both clients and servers are realized as distributed objects communicating through a software bus

37. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 37 Example: A data mining system

38. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 38 Data mining system l The logical model of the system is not one of service provision where there are distinguished data management services l It allows the number of databases that are accessed to be increased without disrupting the system l It allows new types of relationship to be mined by adding new integrator objects

39. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 39 CORBA l CORBA is an international standard for an Object Request Broker - middleware to manage communications between distributed objects l Several implementations of CORBA are available l DCOM is an alternative approach by Microsoft to object request brokers l CORBA has been defined by the Object Management Group

40. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 40 Application structure l Application objects l Standard objects, defined by the OMG, for a specific domain e.g. insurance l Fundamental CORBA services such as directories and security management l Horizontal (i.e. cutting across applications) facilities such as user interface facilities

41. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 41 CORBA application structure

42. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 42 CORBA standards l An object model for application objects A CORBA object is an encapsulation of state with a well-defined, language-neutral interface defined in an IDL (interface definition language) l An object request broker that manages requests for object services l A set of general object services of use to many distributed applications l A set of common components built on top of these services

43. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 43 CORBA objects l CORBA objects are comparable, in principle, to objects in C++ and Java l They MUST have a separate interface definition that is expressed using a common language (IDL) similar to C++ l There is a mapping from this IDL to programming languages (C++, Java, etc.) l Therefore, objects written in different languages can communicate with each other

44. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 44 Object request broker (ORB) l The ORB handles object communications. It knows of all objects in the system and their interfaces l Using an ORB, the calling object binds an IDL stub that defines the interface of the called object l Calling this stub results in calls to the ORB which then calls the required object through a published IDL skeleton that links the interface to the service implementation

45. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 45 ORB-based object communications

46. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 46 Inter-ORB communications l ORBs are not usually separate programs but are a set of objects in a library that are linked with an application when it is developed l ORBs handle communications between objects executing on the same machine l Several ORBS may be available, and each computer in a distributed system will have its own ORB l Inter-ORB communications are used for distributed object calls

47. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 47 Inter-ORB communications

48. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 48 CORBA services l Naming and trading services These allow objects to discover and refer to other objects on the network l Notification services These allow objects to notify other objects that an event has occurred l Transaction services These support atomic transactions and rollback on failure

49. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 49 l Almost all new large systems are distributed systems l Distributed systems support resource sharing, openness, concurrency, scalability, fault tolerance and transparency l Client-server architectures involve services being delivered by servers to programs operating on clients l User interface software always runs on the client and data management on the server Key points

50. ©Ian Sommerville 2000, Tom Dietterich 2001 Slide 50 Key points l In a distributed object architecture, there is no distinction between clients and servers l Distributed object systems require middleware to handle object communications l The CORBA standards are a set of middleware standards that support distributed object architectures