Performance, Scaling And Load Balancing

Front 1

List four things that affects responsiveness

Back 1

1) Server load

2) Network load

3) Reliance on third party services

4) Delays in all software components

Front 2

What is throughput?

Back 2

Rate at which computational work is done.

Front 3

What is the unit for throughput?

Back 3

Transactions per unit time

Front 4

List five things that affect throughput

Back 4

1) Processor speed

2) Storage performance

3) Resources overloading

4) Reliance on OS and network services

5) Reliance on third-party services

Front 5

List three things that a provider should have in terms of availability

Back 5

1) Uptime, ideally 99.99%

2) Consistency

3) Fault tolerance achieved through redundancy of components

Front 6

List the issues relating to performance

Back 6

1) Availability

2) Adaptability

3) Security

4) QoS

5) Reliability

6) Depandability

Front 7

How to improve performance?

Back 7

1) Better network speeds

2) Faster CPU

3) More memory

4) More disk space

5) Less process switching

6) Better design

Helps individual webservers but doesn't scale indefinitely due to physical limits

Front 8

What features should scalability have?

Back 8

1) Ability to grow in size according to need

2) Removal of limits to performance

3) Replicate resources to balance load between them

Front 9

What is vertical scaling?

Back 9

Adding resources to a node, scale-up

Front 10

What is horizontal scaling?

Back 10

Add nodes, scale-out

Front 11

What is a key factor in minimizing cycle time and managing versioning environments?

Back 11

Can be addressed by capturing and managing your infrastructure as code

Front 12

Give examples of automation frameworks that allow for managing infrastructure as code

Back 12

1) Chef

2) Puppet

3) Ansible

4) Saltstack

Front 13

What are most performance problems?

Back 13

Mostly competition for shared resources

Front 14

What does a load balancer do?

Back 14

Allows applications to proceed concurrently without all competing for the same resources

Front 15

Give an example of load balacing

Back 15

JavaScript to validate user form input in browser to avoid repeated round-trips to server

Front 16

Give an example of a server load balancing architechture

Back 16

Involves use of router in every browser request

1) Request goes to router who then passes it on to a server process

2) Scales poorly as the router can be a major bottleneck

Front 17

List several scheduling algorithms for server selection

Back 17

1) Round robin

2) Highest response time

3) Highest uptime

4) Lowest load

5) Match request size to server performance

6) Match request priority to server performance

7) Combination of the above

Front 18

How does Round robin work? (server selection)

Back 18

Applies each successive request to each server in turn

Front 19

How does Highest response time work? (server selection)

Back 19

Based on monitoring server performance

Front 20

How does Highest uptime work? (server selection)

Back 20

Based on monitoring server failures

Front 21

How does Lowest load work? (server selection)

Back 21

Based on monitoring server resource illustration

Front 22

List several scheduling algorithms for request selection

Back 22

1) First come first served

2) Priority queueing

3) Shortest job first

4) Shortest remaining time

Front 23

How does priority queueing work? (Request selection)

Back 23

Maintaining a number of request queues with different priorities

Front 24

How does shortest job first work? (Request selection)

Back 24

Prioritises request with lower resources requirements

Front 25

How does shortest remaining time work? (Request selection)

Back 25

Prioritises request where overall session is closest to competition