Table of Contents

Generic sFlow Collector

The purpose of this connector is to analyze the information received via sFlow packets to give network operators a better understanding of the flows of data crossing the network.

Depending on the range of this connector, incoming data might be filtered and handled in a different way.

About

Version Info

Range Key Features Based on System Impact
1.0.0.x Initial version. - -
1.1.0.x Some of the logic has been moved to a separate Generic sFlow Agent connector. 1.0.0.1 An additional Generic sFlow Agent element will need to be created.
2.0.0.x [SLC Main] - New branch supporting a different way of handling sFlow packets.
- This branch works together with the Generic sFlow Manager connector.
- -
3.0.0.x - Offloads aggregated sFlow data to the Elasticsearch database.
- This branch does not support the latest Elasticsearch changes yet.
2.0.0.x DataMiner Indexing needs to be installed on the DataMiner Agent to use this branch.

Product Info

Range Supported Firmware
1.0.0.x sFlow Version 5
1.1.0.x sFlow Version 5
2.0.0.x sFlow Version 5
3.0.0.x sFlow Version 5

System Info

Range DCF Integration Cassandra Compliant Linked Components Exported Components
1.0.0.x No Yes - -
1.1.0.x No Yes Generic sFlow Agent -
2.0.0.x No Yes Generic sFlow Manager -
3.0.0.x No Yes Generic sFlow Manager -

Configuration

Connections

Smart-Serial Main Connection

This connector uses a smart-serial connection and requires the following input during element creation:

SERIAL CONNECTION:

  • Interface connection:

    • IP address/host: 127.0.0.1
    • Type: UDP
    • IP port: 6343

Initialization 2.0.0.x

Before an sFlow collector element will start processing the sFlow packets, the Agent first needs to be confirmed.

All Agents sending sFlow data to this collector will be listed in the Agents table, but confirming this Agent can only be done via the sFlow Manager element. Once the sFlow Manager confirms this Agent for the collector, the collector will start processing the received packets from the Agent.

How to Use

Only one sFlow Collector element is supported per DataMiner Agent. These elements will listen to port 6343 of the Agent for incoming sFlow packets and process them (depending on filtering).

There are multiple branches of this connector, which all work differently.

1.1.0.x

The Network Devices Overview is the default page. It contains the sFlow Agents Table, which displays all sources from which sFlow packets are received.

The received sFlow data is offloaded to files. On the Configuration page, you can configure how long these offload files should be kept in case they were not processed by the sFlow Agent element.

You can also add or remove entries in the ACL table in order to filter specific sFlow data.

2.0.0.x and 3.0.0.x

Once an sFlow Agent is confirmed by the sFlow Manager element, the collector will start processing the packets from this Agent.

You can configure additional filters to make sure the sFlow Collector only processes the packets that you are interested in. These filters need to be set from the sFlow Manager.

There are some statistics available in the sFlow Collector that can be used to monitor how many packets have been received, how many are processed and if there is any delay in processing them.

The sFlow Collector will aggregate the sFlow packets and keep them in memory for 1 minute (sliding window). To investigate the traffic from the last minute, details about the aggregated sFlow data can be retrieved using the sFlow Manager.

Depending on the branch, the data can also be offloaded.

  • In range 2.0.0.x, data is offloaded to files. However, this should only be done for debugging purposes.
  • In range 3.0.0.x, data is offloaded to the Elasticsearch database. However, this range does not support the latest Elasticsearch implementation yet.