Published on March 23, 2012
PowerPoint Presentation: R eliable A rray of I ndependent N odes IMPS College Of Engg. & Tech. Submitted by:- Manish Sasmal Saurav Kr. Dey Soumya Roy Sourav Kundu Khusboo Anand Sudipta Paul Guided By: Mr. Rahul Karmakar Existing Problems on Internet:: Existing Problems on Internet: Single points of faliures They are devices that have no inherent redundancy or backup. Bottlenecks They are devices that do not have enough processing power to handle the amount of traffic they receive. These two problems hinder the reliability and performance of the network. What is RAIN Technology?: What is RAIN Technology? RAIN Technology origineted at California Institute Of Technology and its purpose was to overcome the existing problems on the internet. A component that stores data across distributed processors and retrieves it even if some of the processors fail. A communications component that creates a redundant network between multiple processors and supports a single, uniform way of connecting to any of the processors. A computing component that automatically recovers and restarts applications if a processor fails. Goals of RAIN Technology:: Goals of RAIN Technology: RAIN technology was able to offer the solution by minimizing the number of nodes in the chain connecting the client and server By RAIN tecnology making the existing nodes more robust and independent of each other RAIN technology provides the novel feature of replacing a faulty node by a healthy one there by avoiding the break in information flow. In effect with the aid of RAIN connection between a client and server can be maintained despite all the existing problems. Architecture: : Architecture: Features:: Features: Communication: Since the network is frequently a single point of faliure, RAIN provides fault tolerence in the network Group Membership: A fundamental part of fault management is identifying which nodes are healthy and participeting. Data Storage: It is achieved through redundent storage schemes like storing data over multiple disk. Communication:: Communication: Bundled Interfaces: Nodes are permited to have multiple interface card which increases fault tolerance and bandwidth. Link Monitoring: A link-state monitoring protocol that provides a consistence history of the link state at each endpoint Fault-Tolerant interconnect Topologies: Network partioning is always a problem when a cluster of computers must act as a whole. Fault-Tolerant interconnect Topologies:: Fault-Tolerant interconnect Topologies: The Problem: Given n switches connected to m nodes in a ring , what is the best way to minimize the possibility of partioning the nodes when failures occour? A naive approach:: A naive approach: In this construction we simply connect the nodes to the nearest switches A ring is 1-fault tolerant so, we can loose 1 switch without upset A 2 nd switch faliure can partion the network. Diameter Solution:: Diameter Solution: Here the nodes are connected to switches that are maximum distance apart from each other which is diameter in a ring. Diameter construction with nodes of degree 2 connected to n switches of degree 4 can tolarate 3 fault without partioning the network which is optimal. Consistent History Protocol for link faliure:: Consistent History Protocol for link faliure: Bounded Slack :It is a simple stable protocol for monitoring connectivity that maintains consistent history with bounded Slack. Stability : Each read channel event will cost at most some bounded number of observable state transition, preferably one at each endpoint. Correctness : The protocol will eventually correctly reflect the true state of the chanel. PowerPoint Presentation: Group Membership B A D C ABCD ABCD ABCD ABCD link/node failures dynamic reconfiguration Critical building block Tolerating faults is a very tedious. Group Membership ensures that all processes maintain a consistent view of global membership. PowerPoint Presentation: Group Membership: The key to fault detection B A C D group membership list sequence number Token carries: Only unicast messages. 1: ABCD Token based group membership protocol Token mechanism 911 mechanism PowerPoint Presentation: B A C D 1: ABCD 1 B A C D 1 2 2: ABCD B A C D 1 2 3 3: ABCD B A C D 1 2 3 4 4: ABCD PowerPoint Presentation: B A C D 5 2 3 4 B A C D 5 2 3 4 Node or link fails: B A C D 5 3 4 Node or link fails: ? B A C D 5 3 4 5: ACD Node or link fails: If a node is inaccessible, it is excluded and bypassed. Aggressive Failure detection PowerPoint Presentation: B A C D 5 6 4 6: ACD B A C D 5 6 7 B A C D 5 6 7 Node with token fails: A C D 5 6 B Node with token fails: ? ? PowerPoint Presentation: A C D 5 6 B Node with token fails: If the token is lost, it is regenerated . A C D 5 6 6: AC B 5: ACD Node with token fails: Highest sequence number prevails . A C D 7 6 B 911 mechanism PowerPoint Presentation: A C 7 6 Node recovers: B D Recovering nodes are added. A C 7 B D 7: ADC Node recovers: A C 7 6 B D 8: ADC 8 Node recovers: A C 7 9 B D 9: ADC 8 Node recovers: Dynamic Scalability Data Storage:: Data Storage: Erasure-correcting Code: Erasure correcting codes are mathematical means of representing data so that lost information can be recovered. With an (n,k) erasure correcting code, we represent k symbols of original data with n symbols of encoded data. With an m erasure correcting code , original data can be recovered even if m symbols of encoded data lost A code is set to be Maximum Distance Seperable(MDS) if m=n-k. The only operation needed for encoding and decoding are exclusive OR(XOR) operations. Data Storage:: Data Storage: Distributed Store/Retrieve Operations: Suppose, we have n nodes for store operation , we encode data of size d into n symbols, each of size d by k We store one symbol per node For retrive operations we collect the symbols from any k nodes and decode them to obtaib original data Features: Original data can be recovered with upto n-k nodes failure. It provides dynamic reconfigurability. PowerPoint Presentation: Original data can be recovered with upto n-k nodes failure. It provides dynamic reconfigurability. Features of Data Storage: PowerPoint Presentation: switch bus network Heterogeneous network of nodes and switches node node node node switch node node RAIN Platform PowerPoint Presentation: A B C D switch Video client & server on node. Network based on available technology: PowerPoint Presentation: A B C D switch Video client & server on node. Network based on available technology: PowerPoint Presentation: RAIN Technology Proof of Concept: Video Server: A B C D switch1 switch2 Video client & server on every node. PowerPoint Presentation: Link Failure: A B C D Switch1 switch2 PowerPoint Presentation: Switch Failure: A C B D switch switch Continuous reconfiguration (e.g., load-balancing ). 1 2 PowerPoint Presentation: Limited Storage: Insufficient storage to replicate all the data on each node. PowerPoint Presentation: High Availabilty Video Server: A B C D switch switch Video client & server on every node. PowerPoint Presentation: SNOW : Strong Network of Webservers(SNOW) uses the distributed state sharing mechanism enable by RAIN System. SNOW system is a cluster of servers network that provide a highly robastic system. PowerPoint Presentation: Advantages: •There is no limit on the size of a RAIN cluster. •There is no concept of master-slave relation. •A RAIN cluster can tolerate multiple mode failurs. •New node can be added into the added into the cluster to participate in load sharing. •A part of cluster can taken down for maintance while other part can contains the work. •It work with many different intermate application. •This is highly efficient in traffic manegment. PowerPoint Presentation: Development of API’s for using the various building blocks. The implementation of a real distributed file system using the partitioning scheme developed here. The Group Communication Protocols are being extended to address more challenging scenarios. Future Scope PowerPoint Presentation: Thank You!