database - DB Design for high amount of data (20 millions rows / day) -
we looking create software receive log files high number of devices. looking around 20 million rows day log (2kb / each each log line).
i have developed lot of software never large quantity of input data. data needs searchable, sortable, groupable source ip, dest ip, alert level etc.
it should combining similiar log entries (occured 6 times etc..)
any ideas , suggestions on type of design, database , general thinking around appreciated.
update:
found presentation, seems similar scenario, thoughts on this? http://skillsmatter.com/podcast/cloud-grid/mongodb-humongous-data-at-server-density
i see couple of things may want consider.
1) message queue - drop log line , let other part (worker) of system take care of when time permits
2) nosql - reddis, mongodb,cassandra
i think real problem in querying data , not in storing.
also need scalable solution. of nosql databases distributed may need that.
Comments
Post a Comment