Cassandra is able to run on several machines and data centers simultaneously, while users see only

Cassandra is able to run on several machines and data centers simultaneously, while users see only a point. In many relational databases distribution exists you must create a server as Master and the other as Slave, and to stop the Master, who is who manages the distribution, a stop occurs. Already on Cassandra does not happen because it is decentralized and each node is identical to the other existing nodes and it uses a P2P protocol to maintain sync data between us and the living dead. So on a Cassandra environment there is no single point of failure for all nodes in a cluster can ho carillon works exactly the same. In relational databases, there are a few options that treats this distribution as the "data replication" but also makes use of the concept master / slave and one of the servers is as the master and the others are replicating data if the master stop work of other kicks in, so getting loopholes for failures. Cassandra in a machine stop no problem as the others that are running can ho carillon are identical and are in full operation.
The elastic scalability that Cassandra data bank uses is a special property of horizontal scalability, so that scale up and down. Scaling up is when you need to increase the power of their hardware machines or even add new machines, with Cassandra you simply put the machine so that the cluster enxerque and ready, it is not necessary to make a copy of all data for this new machine because the cluster itself takes care of it and keep all the machines synchronized, no need to stop, restart, or do any maintenance on your database, so the more machines added more power to you in your database. can ho carillon Scaling down is when you need to remove hardware migration for some reason or something to that effect, with Cassandra nothing need be done since only one node is less but the other is still working for equal.
All computers are not free of faults that can paralyze a database as its unexpected shutdown or even an unplugged network cable. There are many faults that can occur on a machine in operation, with time large companies was creating can ho carillon machines that can withstand or dodge these flaws, but certainly its cost is very high and sometimes you do not get totally free of flaws. If you have a 24/7 application in a particular datacenter with several machines in operation, there is no guarantee of a catastrophic event inactivate your datacenter. With Cassandra you can work in a cluster that can be hosted on several can ho carillon different can ho carillon datacenter so that if a machine or datacenter stop, his bank had not been unavailable, thus giving sufficient time for its correction, and the functionality of the Cassandra distributed, as soon as that machine or datacenter come into operation all the nodes are synchronized and released for new connections.
Consistency means that the bank always returned the most current record contained in their files. Imagine that we are in an environment can ho carillon of electronic commerce and you are viewing an item that someone else also currently viewing, and it is the last item in stock. If you click add to shopping cart This item should not be available can ho carillon for another user. This issue of consistency always generates a lot of controversy regarding the database because each application has its maturity and know what is really important to maintain consistency with Cassandra and set it there and some types such as:
1 Strict Consistency Sometimes this method is called can ho carillon sequential consistency, and is the lowest can ho carillon level of consistency. It requires that you always return the latest value of a written record. This model apparently seems to be great for this bank that works only on a machine that does not have to check other nodes in a cluster, because the data that is here may be that at this moment are not on another machine that is being updated, then the care must be very large.
2 Causal Consistency This is a weaker form of rigid consistency, which does very thorough use of the machine, which does not exist, all the nodes in a cluster while this synchronized, can ho carillon then this form written in sequence analyzes that The previously inserted item is the new casual they are consistent.
3 eventual consistency (weak) This order means that all the nodes will be updated before releasing any action, even if it takes some time. This may be one of the most interesting ways level of consistency but can generate some problems regarding the synchronization failure. To choose the level of consistency to be applied in your bank you should analis