The packet traces consist of time-stamped records of packets, observed in certain passive measurement point. They can be useful for several purposes. First of all, the analysis of packet traces can provide some knowledge about the state and availability of network resources, which is essential for efficient traffic engineering and network dimensioning. Moreover, collecting packet traces collected in two measurement points constitutes a base for the passive method of QoS measurement. Finally, in-depth analysis of collected packet traces is necessary for developing and validating analytical models of traffic at the packet level.
The flow traces contain the records of flows (identified by the pairs of IP addresses and port numbers), observed in certain passive measurement point. Flow-level traces are often collected in operational networks for the purpose of usage-based accounting. Moreover, they are important for the research on modelling the traffic on the flow, or connection level.
- Tools for capturing packet traces are e.g.: netflow, sflow, IPFIX
QoS measurement results
The repositories of QoS results store the values of
metrics of packet transfer quality (i.e. packet delay, packet delay
packet loss ratio) between certain two measurement points. The purpose
measurement is e.g. to evaluate the QoS delivered to the customers of
operational networks, or to perform trials for validation of new QoS
mechanisms. The raw data sets of QoS results consist of collections of
singleton values, obtained for the measurement packets transferred
considered measurement points.
- Tools for QoS measurements are e.g.: CMToolset
The HTTP traces consist of the records of HTTP messages captured in passive measurement point, which is usually located in some operational WWW server. The HTTP traces can be collected for the purpose of assessing the load imposed on the web server by the incoming user requests. Modelling traffic at the session level (where session is related with the user activity, resulting in the exchange of HTTP requests and responses) is considered as a difficult task. The HTTP traces can be essential in the work on new session-level traffic models.
- HTTP traces are usually captured by collecting web server logs.
WEB based repositories
The last type of data in MOME repository is covered by the general category of web-based repositories. It includes for example the collections of measurement data of different types, commonly stored in a single repository.