Introduction

Tom Kelliher, CS43

Feb. 6, 1996

Introduction

Syllabus

1. Contract between students and instructor

2. Keys: Responsibility, Commitment, Discipline

3. Late assignments accepted only with prior approval

4. Use of WWW --- abuse of WWW

5. Perspective on collaboration --- useful aid, poor crutch

6. Working in Unix, Unix Lab, reference books, keys

7. Tentativeness, creation of schedule

What I Expect

1. Demonstration of reasoning, problem solving and decomposition skills

2. Communication, correctness demonstration skills

3. Ability to detect, resolve ambiguity; exercise decision skills

4. ``Fluencies'':
   • To develop an awareness of the issues involved in distributed systems.
   • To achieve an understanding of the design models and principles associated with distributed systems.
   • To gain hands-on experience with distributed systems by configuring and studying client services on Unix workstations and through writing some distributed tools and applications.

5. Seminar/Lab organization; extensive discussions.

6. Proficient in C, comfortable in Unix.

Other information

Need to set-up Unix accounts for those who don't have them.

Introduction to Distributed Systems

What is one?

A distributed system is a collection of independent computers that appear to the users of the system as a single computer.

Historic, Future Trends

• Centralized computing
• Workstations (GUIs, economics, LAN technology)
• Return to centralized? (``big'' pipes, latency, economics, Java)

Distributed System Advantages

Over mainframes

1. Workstations cheaper than mainframes: better price/performance
2. Inherent distribution --- Walmart, groupware
3. Reliability
4. Incremental growth

Over standalone PCs

1. Resource sharing
2. Enhanced communication
3. Flexibility, mobility
4. Load reallocation

Distributed System Disadvantages

1. Software: fully utilizing the resources, hiding the distributed nature of the system
2. Network saturation
3. Network outages
4. Security

Distributed System Hardware

A Taxonomy:

Compare/contrast: message delay, bandwidth, scalability, cost

Bus-Based Multicomputers

• Private memory
• Workstations on a LAN
• Communication Latency

Switched Multicomputers

• Mesh --- dependent routing
• Hypercube --- independent routing

Generalizes to WANs

Software

Loosely Coupled Distributed Software

• Individual machines are distinguishable
• rlogin abacus
• Individual OSs pieced together with a NOS

NOS Functions

• File, print servers (NFS, Netware, printcap)
• DNS, RPC
• User information (YP, NIS, NetInfo, NDS)
• email, ftp, gopher, ...

Tightly Coupled Distributed Software

• True distributed system --- indistinguishable machines
• Unifying OS --- single system image

Characteristics? Global and consistent are keys in the following:

• Namespace (files, services, everything)
• Security
• Communication
• Process management
• Reliability, redundancy

Design Issues for Distributed Systems

Transparency

Hide distribution from users (easier), software (harder)

Types of transparency:

• Location
• Migration
• Replication
• Concurrency --- resource sharing
• Parallelism --- utilize multiple CPUs for a single task

Flexibility

``Keep options open.''

Microkernels

Reliability

Availability, security, fault tolerance

Redundant system should be more reliable than a centralized system

Performance

Communication is the bottleneck

Reliability, transparency affect performance

Fit the job to the topology

Scalability

If it works for 200 CPUs, will it work for 200,000,000?

Avoid centralization of: hardware, software, algorithms (routing info.)

The Internet's anarchy