• For distributed systems • Makes cluster coordination – Fast – Scalable • Runs on a cluster of servers • Offers pre defined functionality • No need to “re invent the wheel” • Best for read dominant work loads www.semtech-solutions.co.nz [email protected]
an ensemble of servers • Clients write changes to ensemble • Changes processed strictly in order received • Ensemble selects a leader • Leader re elected in case of failure • Designed for small data volumes / meta data www.semtech-solutions.co.nz [email protected]
successful if – It reaches a quorum – Quorum is at least half of ensemble • A server will disconnect clients if – It cant contact quorum within a given time • Cluster will function if half of servers up • When a failed server re starts it will – Re sync with the ensemble www.semtech-solutions.co.nz [email protected]
encounter • Inconsistency – Configuration inconsistencies across cluster • Race conditions – Unexpected behaviour due to timing of events • Dead locks – Contention for resources www.semtech-solutions.co.nz [email protected]
name space • Like a file system but for low data volumes • Path elements separated by a slash ( “/” ) • Every node in name space identified by a path www.semtech-solutions.co.nz [email protected]
www.semtech-solutions.co.nz – [email protected] • We offer IT project consultancy • We are happy to hear about your problems • You can just pay for those hours that you need • To solve your problems