Perhaps one of the most popular network services is J-Flow. J-Flow allows you to sample a subset of traffic and collect flow statistics. Flows are identified by unidirectional conversations. A flow is uniquely identified by the following fields:
Source IP address
Destination IP address
Source port number
Destination port number
Type of service
Collecting flow information is critical for businesses to provide accounting and billing, network capacity planning, and traffic profiling and analysis, and some countries require by law that all connections be collected. Flow information is created and transmitted to an external collector for further processing.
J-Flow is used to describe many different variants of collecting flow statistics. Each successive version of J-Flow provides more features and functionality than the previous version.
This version of J-Flow supports only IPv4 and fixed fields that are not user-configurable.
Flow aggregation was added with J-Flow v8. This enables the router to use less bandwidth, sending flow statistics to collectors. Another benefit is that the aggregation reduces the memory requirements of the collectors.
The introduction of RFC 3954 introduced new concepts into flow statistics. The most notable was the introduction of predefined templates such as IPv4, IPv6, MPLS, and IPv4 in MPLS. Templates allow the router and collector to describe the flow fields in a common language. This allows ...