Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- ===unstructured Peer-to-Peer architectures===
- *randomized Algorithms for constructing the overlay Network
- *Each Node maintains a List of neighbors
- *Random Graph, partial View, push and pull Mode
- ===topology Management of overlay Networks===
- *two layers:
- ** Random overlay
- ***nodes Exchange entries periodically, Random graph
- **structured overlay
- ***gets Data from lower Layer
- ***additional information
- ***ranking, ordering by distance, Age, semantic proximity
- ===superpeers===
- *avoid Request flooding
- *CDN use brokers
- *handle all nodes Communication
- *highly available, Backup schemes
- ==Hybrid architectures==
- ===Edge-Server Systems==
- *Hybrid Architecture
- *Clients connect via ISP Edge Servers
- ===collaborative distributed Systems===
- *Bit torrent: p53 F2-14
- ==architectures versus middleware==
- *specific architectural style
- **Object-based like CORBA
- **Event-Based like TIB/Rendezvous
- *drawback middleware-application
- *requirements: easy to configure, adapt, customize
- *Separation between policies and mechanisms
- ===interceptors===
- ;interceptor: Software constructing that Breaks the usual Flow and executes other code
- ==Fault tolerance==
- *a partial failure in a distributed system may affect operation of other components
- *Important Goal: automatically Recovery from partial failures
- ;process resilience: processes can fail without disturbing other Processes
- *reliable multicasting for messages transmission
- *Atomicity
- ===Introduction to fault tolerance===
- ====Basic Concepts====
- #Availability
- #Reliability
Add Comment
Please, Sign In to add comment