MobileStack

To monitor quality of service (QoS) and assure quality of experience (QoE) for a mobile subscriber, you need complete visibility for all of a subscriber’s data sessions. QoS and QoE are monitored with probes that generate relevant key performance indicators (KPIs) based on subscriber traffic. Probes, however, are not designed to correlate session traffic. With their resources dedicated to other functions, a capacity limitation ensues in regards to the number of subscribers, throughput, and packets per second (pps) they can support. With the GTP/SIP session correlation feature, you can recreate a subscriber's full data session by tapping the various control and data plane interfaces and directing all traffic belonging to a given user to the same monitoring probe. Offloading the correlation to Vision ONE, GTP Session Controller (GSC) or CloudLens can free probe resources by up to 50% — allowing you to scale out your monitoring infrastructure.

• True correlation covers at least S1-U, S11 interfaces, as well as handovers from LTE to 4G to 3G, etc.
  
• Other interfaces usually covered by probes are S5/8, S6a, S1-MME

To scale a monitoring infrastructure, load balancing is required to ensure network probes get distributed traffic or sessions. Distribution ensures multiple probes can split the load of monitoring quality of service (QoS) without exceeding a given probe's capacity.

• Enabled by load balancing GTP/SIP data plane traffic based on probe throughput capacity
  
• Enhanced by GTP/SIP session correlation function, assuring that traffic for one subscriber’s control and data sessions always reaches the same monitoring probe

Coupled with GTP/SIP correlation functionality, load balancing can be done on more than just throughput – one can load balance based on the number of supported subscribers per probe or by the supported packets per second (pps), which are often exceeded long before throughput.

Subscriber aware filtering enables two crucial use-cases for operators. First, it can be used as a debugging technique. Data from users who are using new features can be selectively filtered and analyzed. Second, it can be used with the subscriber sampling feature to assure always-on monitoring of high value subscribers.

You can get the same globally relevant statistics without losing visibility of key customers. This allows flexibility in ensuring visibility of user experience by different subscriber segments. It reduces costs associated with infrastructure while allowing an operator to provide customer's different SLAs.

Subscribers can be identified either by IMSI (International Mobile Subscriber Identity) or IMEI (International Mobile Equipment Identity).

Reduce costs by sending only a certain percentage of the subscriber traffic to the monitoring infrastructure. Help improve throughput and limit load on tools, while still getting the insight needed to manage QoS and monitoring quality of experience (QoE).

• If a representative sample of data is selected, the variation in Probe KPIs (e.g. the percentage of calls dropped over a given period) will not be statistically significant.
  
• This feature can be used together with subscriber aware filtering to provide whitelist capabilities for high value customers that are never sampled out.

Do you have a VIP customer that always needs to be included in your sampling? No problem. Ixia offers a whitelist capacity of 8 million customers based on IMSI or IMEI. You don't have to worry about losing visibility for your high value customers as subscriber whitelisting ensures these customers are always included in the session sample sent to probes for monitoring. To obtain the full call for monitoring, you have the option to whitelist by caller or callee so that the cause of an issue can be pinpointed if something were to go wrong.

Reduce monitoring costs by selectively sending traffic to probes based on traffic type. For example, you could send only voice over IP (VoIP)/voice over LTE (VoLTE) traffic or 4G traffic to probes, but not 2G, 3G, etc. This makes the monitoring infrastructure more efficient, allowing faster resolution of issues based on traffic types.

Parameters for filtering include:

• Radio Access Technology (RAT) – 2G, 3G, 4G
  
• Bearer QoS Class Identifier (QCI), which can distinguish for example between regular data, voice and video conferencing
• APN (Access Point Name, which can further be used for differentiating between the data APN and VoLTE IMS for example)