I’m reading many links about High Availability, Elastic Systems and NoSQL alternatives. It’s all about alternatives for scaling a system where a RDBMS can’t fill it needs.
We are putting them below, with some messages and images:
NOSQL Patterns – How we can build a distributed database
This article describes the principal NoSQL alternatives and how to split a table/dataset by key an put them into many servers. It also present how map reduce occurs in the cluster nodes with detailed diagrams.
RoadMap to SaaS
How to convert a standard web application in to a SaaS with Facade class and RESTFul services, moving processes to different JVM’s, cache tiers and much more…
Observers: Making ZooKeeper Scale Even Further
Maybe you don’t know about Zookeeper: yet. But you can be interested in what you can do to scale servers that depends on distributed communication and synchronization. That is a big trouble: How much more computers you put in the cluster, more messages need to be send to coordinate the distributed lock between the computers. The throughput obviously decreases. That is not true for this guy’s using followers and observers!