principles of engineering system design

Principles of Principles of

Engineering System DesignEngineering System Design

Dr T Asokan

asok@iitm.ac.in

INTRODUCTION TO SYSTEMS DESIGN

Dr T Asokanasok@iitm.ac.in

Interface Architecture Development

Six functions of Design Process

1. Define System Level Design Problem :- Originating requirements development

2. Develop the system functional architecture

3. Develop the system physical architecture

4. Develop the system operational architecture

5. Develop the interface architecture

6. Define the qualification system for the system

T Asokan ED309

CASE STUDY: PATH FINDER

• Path finder system that was deployed to the surface of Mars for landing on 4th July 1997 was a great success in many ways.

• Few days into the mission, operators on the ground noticed that total system resets were occurring that were causing the loss of data.

• A shared memory interface was used as the system interface between various sub

systems.

• Mutual exclusion (mutex) locks were employed to give an activity access to the interface.

CASE STUDY: PATH FINDER

• Meteorological data was so voluminous that the activity had to obtain and release mutexes several times before it was finished.

• The long running, medium priority, communication activity would infrequently interrupt the meteorological activity during its pause and gain control of the interface.

• Duration of these two tasks were sufficiently long to invoke a watchdog timer that was employed to ensure that the high priority bus management task was executing

appropriately.

• In such rare cases, watch dog timer initiated a total system reset to prevent any further damage to the system.

• Jet Propulsion Lab. Engineers ran a pathfinder replica on earth till they reached the reset situation.

• It was found that the interface software (VXworks) had been programmed without a feature called ‘priority inheritance’.

• JPL engineers uploaded a short C programme and Pathfinder did not experience anymore system resets.

INTERFACE DESIGNInterfaces are common failure points in the system.

An Interface is a connection resource for hooking to another system’s interface (an external interface) or for hooking one system’s component to another (an internal interface).

• Identifying interfaces (external and internal)

• Allocating inputs and outputs for the interfaces

• Derive interface requirements

• Exploring alternative interface architectures

T Asokan ED309

Interface requirements

• Performance: Throughput and response time

• Fidelity ( should not change the data during transmission)

• Deliver every item placed on the interface

• Should detect faults and recover gracefully

T Asokan ED309

Generic Interface architectures

MESSAGE PASSING: - mail delivery that predictably occurs once or twice a day, allowing the receivers to access it immediately or wait until a more opportune time

SHARED MEMORY : A meeting or conference in which only one person speaks and conveys relatively compact messages, all can hear what is said but yet are restrained from other productive work.

NETWORK: a telephone conversation that can involve messages of widely varying lengths and can be instigated at almost any time.

T Asokan ED309

Message Passing architectures

• Used to allow predictable exchange of information

• Commonly found as an internal interface

• Message consists of a protocol and data segments

• Protocol segment includes:

• size of the message

• address of the node to receive the message

T Asokan ED309

Process of communication

1. One node must win control over the communication channel by a priority scheme implemented by the system

2. The winning node becomes the master and sends a protocol segment to the intended receiving node(s), called slave(s).

3. The slave node notifies the master that the protocol segment was received successfully.

4. The master sends/receives the data segment to/from the slave.

5. The slave notifies that the data segment transfer is complete

6. The master surrenders control of the communication channel.

Most preferred application of message passing is for systems that can define a predictable message transmission schedule upon initialisation

Updates rates are on the order of 0.01 to 1 second

Message passing is not preferred where substantial portions of the traffic include asynchronous communication

T Asokan ED309

Shared Memory Architectures

• Asynchronous communication requests are handled

• A fast access storage device, typically a memory device

Communication process

• A processor generates a read or write request for another address in shared memory

• The current owner of this variable is notified of the request

• The cache memory of the current owner is dumped to shared memory

• The read or write request of the processor is completed with a data transfer

Performance of shared memory degrades substantially if the requested information is not in the cache memory of the interface

In this case all activity is blocked until the required variables are retrieved

Works best in highly parallel software applications in which the global data of each application must be accessed frequently by the application and infrequently or never by other applications

Network Architecture

• A distributed collection of shared memory systems : LAN

• Each shared memory system has the ability to tap into the shared memory of other systems

• Provides a demand-based service, unlike message passing where scheduled transfers take place.

• Networks can serve hundreds of nodes (message passing architecture is limited to 32 nodes)

• Includes communication hardware and software package, called network operating system

• Software provides priority based queuing models

• Provides extensive fault checking

Most common types of Network Architectures :

• Master-slave or pipeline

• Bus

• Star or Spoke

• RING

• MESH

C1

C2

I12

C3

I23

Pipeline architecture

•Appropriate when the components only need to communicate with their neighbor in the network

I123

C1 C2

C3

BUS architecture

•Most common architecture•Appropriate for large number of components

C1

C2

I12

C3

I34

C4

I24

Star or Spoke Architecture

• Isolates one component as central processor that manages the communication

C1

C2

C3

I12 I23

I13

RING Architecture

•Suitable for office settings

C1

C2

C3

I12 I23

I13 C4

C7 C6 C5

I13

I67

I17

I56

I45I46

MESH Architecture

•Provides redundancy•Used in parallel computing and telephone networks