By Stephane Bourque
Service providers today face increasing demand for bandwidth from consumers using bandwidth-hungry applications and services. Video content delivered via over-the-top (OTT) providers like Netflix can drain a huge amount of bandwidth, and this type of service is growing in popularity worldwide. Some MSOs have turned to usage-based billing models to stem the tide, but it is difficult to make informed decisions about OTT strategy unless you understand what is happening in your network. Proper network intelligence can tell you not only where subscriber bandwidth is used, but also where congestion occurs, and at what time traffic peaks. This data can then be used to create solutions that maximize service quality. Network intelligence is therefore essential to enhance your network, develop faster and more reliable service, and ultimately, improve the subscriber experience.
Find the Hidden Information in Your NetworkThe first step is to find out exactly where your bandwidth is going. Proper, controlled and predictable system upgrades can help you extract every bit of traffic left on the network. In the past, there was a lack of available resources to properly analyze network usage and plan for growth, but software vendors are now creating a new breed of network-intelligence tools that use a range of data sources to enable providers to work smarter and deliver bandwidth where it is needed. Some of these data sources include Simple Network Management Protocol (SNMP), IP Data Record (IPDR) and sFlow. By themselves, each of these data sources provides valuable quantitative and qualitative information. But analyzing and cross-referencing these data sources with each other allows you to correlate the results into something much more meaningful, which can lead to better and faster decision making.
Simple Network Management Protocol (SNMP) data is the grandfather of it all and provides important information about network health. This protocol has been used in the wireline and cable industries for decades and allows the extraction of data from most networking equipment. This includes quantitative measures such as the number of bytes or packets sent or received through any given interface on the network, and qualitative data such as the number of dropped packets, retransmission, noise levels and so on.
IP Data Record (IPDR) is the equivalent of Call Detail Record (CDR) in the telephony world. This technology is integrated into DOCSIS and has been used almost exclusively by cable in the past, but more recently has been put forward as an option for telecom. IPDR focuses specifically on subscriber usage. In cable, for instance, IPDR collects information from every flow inside a CMTS and provides consumption information about subscriber devices on a network. This data provides a way for carriers to represent subscriber usage and analyze traffic patterns in specific regions.