Nave dsm assume each host has a local copy of all of

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: have all replies: #1: All other processes are aware of your request #2: You are aware of any earlier requests for the mutex A shared priority queue •  FIFO fairness. •  Client Delay: 3(N- 1) –  (N- 1 requests + N- 1 ack + N- 1 release) •  Very unreliable - any process failure halts progress Ricart & Agrawala’s algorithm •  An improved version of Lamport’s shared priority queue 1. Broadcast a \mestamped request to all. 2. Upon receiving a request, send ack if You do not want to enter your CS, or You are trying to enter your CS, but your \mestamp is higher than that of the sender. (If you are in CS, then buffer the request) 3. Enter CS, when you receive ack from all. 4. Upon exit from CS, send ack to each pending request before making a new request. (No release message is necessary. No priority queue needed) Ricart & Agrawala’s algorithm •  Safety: Prove that at most one process can be in CS. •  Fairness: Prove that FIFO fairness holds even if channels are not FIFO •  Not reliable •  Message complexity = 2(N- 1) (N- 1 requests + N- 1 acks - no release message) Majority rules •  Instead of collec\ng REPLYs, collect VOTEs •  Can progress with as many as N/2 – 1 failed processes •  Not fair •  Deadlock! No guarantee that anyone receives a majority of votes DFS •  Networked file systems with a single- server architecture •  Their goals: –  Have a consistent namespace for files across computers –  Let authorized users access their files from any computer –  That’s why you have the iden\cal files in all CLIC machines. SUN NFS • NFS provides transparent, remote file access • Simple, portable, really popular – (it’s go•en a li•le more complex over \me) • Weak consistency seman\cs • Requires hevy server resources to scale (flush- onclose, server queried for lots of opera\ons) SUN NFS •  NFS V1- V3 are stateless •  NFS V4 is stateful AFS •  Developed at CMU (~1985), commercial product by Transarc (part of OSF/DCE) •  Segment network into clusters, with one file server per cluster •  Dynamic reconfigura\ons to balance load •  Stateful protocol + aggressive caching •  Servers par\cipate in client cache management •  En\re files are cached on disk File System 1. Machine failures are the norm – 1000s of components – Bugs, human errors, failures of memory, disk, connectors, networking, and power supplies – Monitoring, error detec\on, fault tolerance, automa\c recovery must be integral parts of a design 2. Design for big- data workloads – Search, ads, Web analy\cs, Map...
View Full Document

Ask a homework question - tutors are online