Introduction to Information Technology
ITEC 1000 – Winter 2010 – Peter Khaiter
Lecture 11 – Computer Systems and Networks – Mar 22
System Performance Improvements
- Multiple CPUs
o Sharing memory and I/O facilities
o Conflict among the CPUs for shared resources
- Faster clock speed, buses and circuits
o Improving technology to design faster CPU circuits and buses
- Wider instruction and data paths
o Wider interface between the CPU and memory allows to fetch more data
in a single operation
- Faster disk access
o Smaller discs, more density packed, increased storage
o RAID: data on different discs, multiple access simultaneously
- More and faster memory
o Reduces the time to access instructions and data
Multiprocessing
- Reasons
o Increase the processing power of a system
o Parallel processing
- Types of multiprocessor systems
o Tightly coupled systems
o Loosely coupled systems
Tightly Coupled Systems
- Also called multiprocessor systems
- Identical access to programs, data, shared memory, I/O, etc.
- Easily extends multi-tasking, and redundant program execution
- Two ways to configure
o Master-slave multiprocessing
o Symmetrical multiprocessing (SMP)
Master-Slave Multiprocessing
- Master CPU
o Manages the system
o Controls all resources and scheduling
o Assigns tasks to slave CPUs
- Advantages
o Simplicity
o Protection of system and data
- Disadvantages o Master CPU becomes a bottleneck
o Reliability issues – if master CPU fails entire system fails
Symmetrical Multiprocessing
- Each CPU has equal access to resources
- Each CPU determines what to run using a standard algorithm
- Disadvantages
o Resource conflicts – memory, i/o, etc.
o Complex implementation
- Advantages
o High reliability
o Fault tolerant support is straightforward
o Balanced workload
Loosely Coupled Systems
- Clusters or multi-computer systems
- Each system has its own CPU, memory, and I/O facilities
- Each system is known as a node of the cluster
- Advantages
o Fault-tolerant, scalable, well balanced, distance is not an issue
- Two ways to configure
o Shared-nothing model
o Shared-disk model
Shared-Nothing Model
- High speed link between nodes
- No sharing of resources
- Partitioning of work through division of data
- Advantage
o Reduced communication between nodes
- Disadvantage
o Can result in inefficient division of work
Shared-Disk Model
- High speed link between nodes
- Disk drives are shared between nodes
- Advantage
o Better load balancing
- Disadvantage
o Complex software required for transactional processing (lock, commit
phases)
Multi-Core Processors
- Multi-core processor - two or more independent actual processors ("cores")
- Each core can execute cpu instructions at the time (parallel computing)
- Dual-core processor - two cores (e.g. Intel core duo) - Quad-core processor - four cores (e.g. Amd phenom ii x4
- Hexa-core processor - six cores (e.g. Amd phenom ii x6, intel core i7 extreme
edition 980x)
- Octa-core processor - eight cores (e.g. Amd fx-8150)
- Coupling of cores: tightly or loosely
Computer Interconnection
- Communication channel – pathway for data movement between computers
- Point-to-Point connectivity
o Communication channel that passes data directly between two computers
o Serial connection
o Telephone modem
o Terminal controller – handles multiple point-to-point connections for a host
computer
- Multipoint connectivity
o Multidrop channel or shared communication channel
Client-Server Architecture
- Computer server provides services
o File storage, databases, printing services, login services, web services
- Client computers
o Execute programs in its own memory
o Access files either locally or can request files from a server
Beowulf Clusters
- Simple and highly configurable
- Low cost
- Networked
o Computers connected to one another by a private Ethernet network
o Connection to an external network is through a single gateway computer
- Configuration
o COTS – Commodity-off-the-shelf components such as inexpensive
computers
o Blade components – computers mounted on a motherboard that are
plugged into connectors on a rack
o Either shared-disk or shared-nothing model
LAN Topology
- Arrangement of workstations in a shared medium environment
- Logical arrangement (data flow)
- Physical arrangement (cabling scheme)
LAN
More
Less
