Mobile data demand is growing more quickly than operators can deploy new network capacity and mobile operators need to use policy control applications to manage data traffic and ensure subscribers receive appropriate quality of experience (QoE). QoE is a more subjective measure than quality of service (QoS) and requires a more sophisticated approach to policy control and quality measurement and analysis.  To ensure adequate QoE for subscribers, operators need to do more than just implement basic policy enforcement and traffic management.


In this blog I am reviewing a proof of concept from Advantech and Qosmos that extends a deep packet inspection (DPI) platform, used to implement the Policy and Charging Enforcement Function (PCEF),  to also analyze the quality of video passing through the network to subscribers. Advantech is a Premier member of the Intel® Intelligent Systems Alliance. Qosmos is an Affiliate member of the Alliance.. The 200-plus members of the Alliance collaborate closely with Intel® to create hardware, software, tools, and services to help speed intelligent systems  to market.


Mobile Traffic Growth and Quality of Experience

Mobile data traffic is growing rapidly driven by new subscribers acquiring smart phones and existing smartphone users consuming more network bandwidth. Video is the key application driving much of this growth and mobile video traffic exceeded 50% of data traffic for the first time in 2012. The Cisco VNI Mobile Forecast from February 2013 shows mobile device Internet usage growing by an average of 66% per year from 2012 to 2017 to reach 11.2 Exabytes per month.


To meet the growing demand for data traffic Mobile operators are deploying a mix of 3G and LTE services. The PCEF implements operator policies in the 3G and LTE packet core. As mobile bandwidth grows, the PCEF becomes a bottleneck requiring high-performance deep packet inspection and support for 40Gbit/s or greater packet processing. The Intel® Xeon® Processors E5-2600 and E5-2400 Series and Intel® Data Plane Development Kit (Intel® DPDK) are being widely used to implement the PCEF and other elements in the3G and LTE packet core.


Video, voice, gaming and browsing require different network characteristics. By implementing policies that are application-specific mobile operators can increase quality of experience without necessarily using more network capacity. The quality of Video, for example, is dependent on frame rate, delay, jitter, size and codec. By extracting these parameters from the packet stream and analyzing video quality mobile operators can implement policy to ensure appropriate user quality of experience when playing video.

Leveraging DPI for Quality of Experience

Figure 1 shows the Qosmos ixEngine®. The DPI engine is built on top of the Intel® DPDK and extracts metadata and content from the packets flowing through the network. The Qosmos ixEngine® can be used with any Intel multicore processor supported by the Intel® DPDK including the Intel® Xeon® Processors E5-2600 and E5-2400 Series.


The Qosmos ixEngine® supports real-time layer 7 IP flow analysis and identification of protocols and applications based on flow passing and statistical analysis. The Qosmos ixEngine® will extract over 6,000 types of information for up to 1,000 protocols. For video the parameters extracted include frame rate, delay, jitter, size and codec.

Qosmos DPI Engine.png

Figure 1. Qosmos ixEngine®: DPI and Metadata Engine.


The metadata extracted by the Qosmos ixEngine® can be used in a policy control platform to implement application specific policy. Additional content extracted from the packet stream can be used to further analyze application-specific performance. Qosmos has worked with Argon Design and Advantech to develop a proof of concept that uses data extracted by the Qosmos ixEngine® to analyze video quality in the network.


DPI and Video Analysis Applied to Mobile Video Streaming

Figure 2 shows a video analysis demonstration using the Advantech FWA-6510 network appliance and Qosmos ixEngine. The Advantech FWA-6500 acts as a packet generator streaming canned Internet traffic and running a video streaming server for live viewing. Users can browse the server and stream videos through a wireless network connected to an Advantech FWA-6510. The Advantech FWA-6510 is running the Qosmos ixEngine and Argon Design qualmon to capture and analyze packets in real time. The video quality results are viewed through a web browser.

Video Analysis Demonstration using Advantech FWA-6510 and Qosmos ixEngine.png

Figure 2. Video Analysis Demonstration using Advantech FWA-6510  and Qosmos ixEngine.


The Argon Design qualmon recognizes the video streams, and extracts details of the stream and displays them. It also extracts a representative thumbnail from the video stream and analyses the data to assess visual quality and predict when the user device will be starved of data and the user will experience video freeze. This information can then be used to drive application and user specific policy enforcement.


Scalable Solution

The initial proof of concept has been developed using the Advantech FWA-6510 Network Appliance with dual Intel® Xeon® Processors E5-2600 Series and up to eight 10GbE interfaces. Larger systems can be built using an AdvancedTCA (ATCA) platform with up to 12 ATCA blades each integrating dual Intel® Xeon® Processors E5-2600 Series. The Advantech FWA-6510 Network Appliance and Advantech ATCA Blade are shown in Figure 3.

Advantech FWA-6510 and MIC-5332 ATCA Blade Cropped.png

Figure 3. Advantech FWA-6510 Network Appliance and ATCA Blade.




Solutions in this blog:


Related topics:


Advantech is a Premier member of the Intel® Intelligent Systems Alliance. Qosmos is an Affiliate member of the Alliance.


Simon Stanley

Roving Reporter (Intel® Contractor), Intel® Embedded Alliance

Principal Consultant, Earlswood Marketing

Follow me on Twitter: @simon_stanley