Docsity
Log inSign up
Docsity
Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity

Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan

Guidelines and tips
Guidelines and tips
Sell on Docsity
Sell on Docsity
Docsity AI
Docsity AI
Log inSign up

Prepare for your exams

Study with the several resources on Docsity

Find documents

Prepare for your exams with the study notes shared by other students like you on Docsity

Search for your university

Find the specific documents for your university's exams

Docsity AINEW

Summarize your documents, ask them questions, convert them into quizzes and concept maps

Explore questions

Clear up your doubts by reading the answers to questions asked by your fellow students

Earn points to download

Earn points by helping other students or get them with a premium plan

Share documents
download point icon20 Points

For each uploaded document

Answer questions
download point icon5 Points

For each given answer (max 1 per day)

All the ways to get free points
Get points immediately

Choose a premium plan with all the points you need

Study Opportunities

Choose your next study program

Get in touch with the best universities in the world. Search through thousands of universities and official partners

Community

Ask the community

Ask the community for help and clear up your study doubts

Free resources

Our save-the-student-ebooks!

Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors

Load Balancing in Distributed Systems: Exam Questions and Answers, Exams of Information Technology

Miami University - OxfordInformation Technology

A set of exam questions and answers related to load balancing in distributed systems. It covers topics such as static and dynamic load balancing, round-robin, randomized, and multiple-choice load balancing. The document also discusses challenges and solutions related to load balancing, such as herd behavior and g-bounded errors. It is useful for students studying distributed systems and network traffic management, offering insights into the practical application and theoretical underpinnings of load balancing techniques. The content is presented in a question-and-answer format, making it easy to review and understand key concepts.

Typology: Exams

2025/2026

Available from 12/19/2025

Matress
Matress 🇺🇸

3K documents

1 / 9

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
DISTRIBUTED SYSTEMS – LOAD BALANCING EXAM
QUESTIONS AND CORRECT ANSWERS (LATEST
UPDATED)
What is load balancing?
Evenly balancing the load over servers
Minimise the maximum load on any server
- Max load = longest job queue in the system
- Good indicator of balance
3 load balancing questions
I'm hoping these questions are covered later in the slides, but just in case they
aren't...
Delete if covered:
What information is needed by load balancers?
What are the advantages and disadvantages of common load balancers?
Quality and robustness of randomised load balancers?
When might load balancing be applied?
- Popular websites or databases that employ multiple server requests
- Network traffic management over multiple routes to avoid congestion
- Master-worker applications in parallel computing-
Collision avoidance in hash tables
How does the 'balls into bins' process work with Distributed Systems?
Servers are visualised as bins
Tasks (or processes or something) are visualised as balls
Ideally, we want to give the bin with the fewest balls the next balls.
Note: This guys slides suck, i dont know if he wants us to rewatch his lectures
pf3
pf4
pf5
pf8
pf9

Partial preview of the text

Download Load Balancing in Distributed Systems: Exam Questions and Answers and more Exams Information Technology in PDF only on Docsity!

DISTRIBUTED SYSTEMS – LOAD BALANCING EXAM

QUESTIONS AND CORRECT ANSWERS (LATEST

UPDATED)

What is load balancing? Evenly balancing the load over servers Minimise the maximum load on any server

  • Max load = longest job queue in the system
  • Good indicator of balance 3 load balancing questions I'm hoping these questions are covered later in the slides, but just in case they aren't... Delete if covered: What information is needed by load balancers? What are the advantages and disadvantages of common load balancers? Quality and robustness of randomised load balancers? When might load balancing be applied?
  • Popular websites or databases that employ multiple server requests
  • Network traffic management over multiple routes to avoid congestion
  • Master-worker applications in parallel computing- Collision avoidance in hash tables How does the 'balls into bins' process work with Distributed Systems? Servers are visualised as bins Tasks (or processes or something) are visualised as balls Ideally, we want to give the bin with the fewest balls the next balls. Note: This guys slides suck, i dont know if he wants us to rewatch his lectures

or some bs, but compare notes with people for this in particular (setup study sessions?). This card was made with assumptions and not slide knowledge What are some challenges with the balls in bins method? Complexity of procedure Outdated/inaccurate info Fault tolerance Scalability (dynamic system may change size) Heterogeneous hardware (some servers have more capacity or higher speeds) He didn't mention it, so i might be wrong, but i think balls can also take different amounts of time to process so its hard to predict? Maybe thats 'inaccurate info' What is static load balancing? Load balancer does NOT consider system state

  • Current load level not taken into account
  • Everything set in advance (no state info required during execution) What are the applications of static load balancing? Static systems that do not change Tasks that need to be fairly regular What are some strategies for static load balancing? Round-robin Randomised How does round-robin load balancing work? Assign next ball to next bin in sequence Once end is reached, start from beginning How can we deal with heterogeneous hardware with a round robin? Use weighted round robin So if a server had double speed/capacity Give 2 balls Note: This is where server has extra capacity/speed How does round robin work with DNS

Hash function

  • A good hash function distributes items/requests like a random function
  • An item can be found by every client knowing the item ID and hash function What are some pros & cons of random/hash-key load balancing? There is no single point of failure BUT Load is not perfectly balanced What is dynamic load balancing? Decisions based on current state (e.g. load level of bins) Tasks can be moved from one node to another Allows algorithms to be more modular Give some examples of a dynamic load balancer Master-worker Work stealing Least load Least time Randomised with multiple choice How does a least load/connections dynamic load balancer work? Server-sided recevies request and forwards to backend servers Perfect choice Backend servers are NOT directly accessed Commonly used in web servers Note: Only a perfect choice if load balancer is up to date, but i think this is a requirement as its a must on slides What are some pros and cons of least load? Load balancer must be kept up to date Single point of failure (though can be avoided) pros are perfect choice and i think, backend not directly accessed How does the least time dynamic load balancer work?

Similar to least load, but we also consider server speed Computes how long request would take at every backend server given current speed and queues Note: works great for heterogeneous hardware and is applied by Nginx and other web servers What is the problem with a centralised load balancer? It is a potential bottleneck Its a single point of failure We can use multiple balancers to fix this (multiple dispatchers) What is a problem with multiple dispatchers? As they are working in parallel on same info, they can send their requests to the same server (e.g. lowest load) This herd behaviour can have the opposite effect What are the solutions for herd behaviour in multiple dispatchers? Randomised load balancing OR NOTE: - If you need this i will be disgusted Round-based alg with dispatchers allocating jobs to bins according to their probabilities in the current round. To compute probabilities, some info is needed about server queues and processing rates as well as job arrival rates What is multiple-choice randomised load balancing? Dynamic randomised load balancing Instead of choosing 1 random bin (or server), choose multiple random bins Allocate ball to the bin of lowest load among these choices What is the d-choice protocol (greedy[d])? Greedy[d] Allocate m balls into n bins

What is a g-bounded error? Balls allocated using 2 choices, but if load difference is g or less, the process makes a mistake and chooses the higher loaded bin This models scenarios where load queries are inaccurate What is the result of a process being g-bounded? Do i need to know this Hope not This is called saving load queries, but the slide doesn't make much sense right now. Make flash later Note: Stored at tmp In a 1-choice load balancer, what is the probability that a bin is empty? Assume m = n Probability of particular ball going into particular bin: 1/n Probability of particular ball not falling into particular bin = 1 - (1/n) Probability that none of the m = n balls falls into a particular bin for (n → ∞):

  • (1 - (1/n))^m = (1-(1/n))^n = 1/e Note: All these assume m = n, but maybe consider what if this isnt the case just incase its a dumb question. Also try and truly understand e. If n is 5 or n is 10, the result of the formula is different, so how can we use 1/e unless e itself changes? I'm guessing e is the number e, but if so, this formula doesnt match up? In a 1-choice load balancer, what is the expected number of empty bins? Assume m = n Since each bin has a 1/e chance of being empty AND We have n bins We can use n/e to determine the number of empty bins (roughly 1/3) In a 1-choice load balancer, what is the expected number of collisions?

A collision is 2 balls falling into same bin Three balls in a bin count as 3 collisions (1-2, 1-3, 2-3) (m C 2) · 1/n <- This Note: Formula to left, hopefully dont have to understand but check In a 1-choice load balancer, what is probability of a particular bin having exactly k balls? More nonsense i hope i dont have to know (but check) In a 1-choice load balancer, what is probability of a particular bin having at least k balls? Ew In a 1-choice load balancer, what is maximum load? 1 - o(n) He has 2 proof slides on the powerpoint but i cba, I dont think we will have to know them or even understand them. They are confusing and look like a waste of time What is a distributed hash table? Give an example A set of nodes that store data items When data item is added, a hash function is used to determine node. E.g. a chord How can we add a new (actual) node in a chord? let y be successor of x move items at y that belong to node x to x

  • e.g. new node 16 would get all elements from 18 that are hashed to 15 and 16 update finger tables How can we remove an actual node x from a chord? Let y be successor of x Move items of x to y Update finger tables What are the advantages and disadvantages of a chord? Advantages: