Download Distributed Systems: Cloud Computing, Load Balancing, and Concurrency Control UPDATED A and more Exams Medicine in PDF only on Docsity!
Distributed Systems: Cloud Computing, Load
Balancing, and Concurrency Control UPDATED
ACTUAL Questions and CORRECT Aswers
What is the primary difference Traditional architecture uses physical servers and between traditional and modern manual management, while modern architecture data center architecture? uses virtual machines and increased automation. What is virtualization? A technology that creates virtual representations of servers and physical machines, allowing multiple virtual machines to run on a single physical machine.
What are the advantages of Faster hardware changes, ease of portability, code virtualization? migration, and isolation of failing components. What is a virtual machine (VM)? A VM is a software emulation of a physical computer that can run its own operating system and applications. What is Infrastructure-as-a-Service A cloud computing model where a provider rents (IaaS)? out virtual machines and resources instead of physical machines. What is the role of VMware in VMware provides products like VMware ESXi that virtualization? allow VMs to access the resources of physical servers. What is the purpose of load To distribute workloads evenly across servers to balancing in distributed systems? minimize delays and improve performance. What factors should be considered Server capabilities, service requirements, system in load balancing? state knowledge, delay minimization, energy consumption, and request acceptance.
What is the role of a central It collects server load information centrally and coordinator in load information globally to aid in load balancing decisions. collection? What is the purpose of server To determine suitable servers for receiving tasks allocation in load distribution? and to organize collaborative server efforts for load sharing. What is the impact of application- Factors like the number of transactions, users, and dependent factors on workload? media content can affect the workload and how tasks are distributed. What does it mean for a VM to have A public IP address allows the VM to communicate a public and private IP address? with the internet, while a private IP address is used for communication within a private network. What is the importance of isolation in It ensures that failing or attacked components do virtualization? not affect other virtual machines running on the same physical hardware. What is the concept of task It involves reallocating ongoing tasks to balance redistribution among servers? the load across servers effectively. What is the advantage of running Each VM can be assigned one or more virtual CPUs multiple VMs on multi-core systems? (vCPUs), optimizing resource utilization.
What is the difference between Centralized collection gathers data from all servers centralized and local load globally, while local collection gathers data from information collection? neighboring servers. What are privileged instructions? Instructions that can only be executed by the operating system, ensuring security and control. What is the role of system calls in System calls provide an interface for applications operating systems? to request services from the operating system. What is the significance of flexibility It allows for dynamic resource allocation and in modern data centers? scaling based on demand. What is the impact of server Overloaded servers can negate the benefits of overload on replicated servers? replication, leading to performance issues. What is the least-workload policy in It sends the job to the server with the least load balancing? workload, measured by the time to finish currently assigned jobs. What are the exact processing times Knowing the exact processing time of every job in workload estimation? and the remaining service-time of ongoing jobs. How can probability distributions be By knowing the probability distributions of service used in estimating server workload? times, which can be estimated from past data. What does the static approach in It uses a fixed policy where scheduling decisions load distribution entail? do not depend on the current state of the system.
What are the drawbacks of the It can become unstable at high loads due to sender-initiated algorithm? increased polling activity and difficulty in finding receivers. What is a symmetric load distribution Both senders and receivers search for each other, approach? which can lead to instability at high loads. What is the main idea behind an To limit sender's polling actions at high load to adaptive algorithm in load avoid instability while proactively tracking server distribution? status. What is the role of the Information It defines what information is required to support Policy in load distribution? decision-making and how to maintain it. What does the Selection Policy It determines which tasks and which servers will be determine in load distribution? involved in the load distribution process. What is the effect of polling in a Polling initiated by the receiver can make it difficult receiver-initiated approach? to find senders for new tasks, potentially leading to higher transfer costs. What is the significance of the It reduces overhead by limiting the number of PollLimit in load distribution? servers polled during task transfers.
What is the purpose of using Thresholds help determine when a server should thresholds in transfer policies? act as a sender or receiver based on its load. How does the adaptive algorithm Each server maintains lists of all servers, initially maintain server status? assuming they are all receivers, and updates their status through polling. What is the impact of high system High system load can increase polling activity, load on polling activity? potentially leading to system instability. What is the benefit of using a It allows for task transfers without needing random selection in location extensive state collection, although it may lead to policies? unnecessary transfers. What is a preemptive task transfer? It occurs when a lower priority task is moved from a server to accommodate a newly arrived higher priority task. What is the role of the queue length Queue length is used to determine whether a in load distribution? server should be classified as a sender or receiver. What is the challenge of It may inherit disadvantages from both sender- implementing a symmetric load initiated and receiver-initiated approaches, distribution algorithm? especially under high loads. What is the primary function of a To distribute incoming tasks among multiple load balancer? servers to optimize resource use and minimize response time. What is the Power-of-d load A policy that assigns a job to the server with the balancing policy? least queue size out of 'd' servers chosen uniformly at random.
What is the role of a server To ensure the objects remain consistent when managing shared objects? accessed by concurrent events. What is the impact of high loads on At high loads, receiver-initiated algorithms work sender-initiated and receiver- better, while sender-initiated algorithms are more initiated algorithms? effective at low loads. What is an adaptive algorithm in load An algorithm that adjusts its approach based on balancing? the current load conditions. What is the challenge of analyzing The complex dependencies introduced by the systems with a small number of power-of-d policy make analysis difficult. servers? What is the recommended approach Use symmetric algorithms for load balancing. for systems with widely fluctuating loads? What is the main focus of the To efficiently distribute tasks among a large discussed load balancing number of servers, such as in modern data centers. algorithms? What is the benefit of using mean- They allow for accurate analysis when the number field techniques in analyzing large of servers is large, typically over 100. server systems? What is the consequence of a delay Clients may receive outdated information if they in updating data at a server? access the data before it is updated. What is the effect of server removal It gradually reduces sender-initiated polling as from the receiver list during high servers become senders. loads?
