How Application Awareness Drives Intelligent Mobile Data Offloading Decisions

5G holds many promises, including unlimited bandwidth and unrivaled speeds. Today’s mobile users expect 5G to seamlessly deliver rich and highly interactive content to their devices – be it HD video streaming, online gaming, or a virtual reality application – while they move around their familiar surroundings and beyond.

Why HetNets?

In the context of a heterogeneous network (HetNet), in which an operator operates a network of networks, 5G often works in parallel with other last mile access technologies. Several access technologies are bridged in HetNets in order to create deeper and wider network coverage. Access networks can be in the form of macro cells (3G, 4G and 5G radio), small cells (pico, femto and micro cells), distributed antenna systems (DAS) and operator Wi-Fi access points. The traffic from these various access networks converges at local aggregation points before being routed back to the core and routed to the Internet.

In HetNets, the access networks essentially overlap. This enables traffic shifting, a process in which the operator routes user traffic from one access point to another on a different access network based on network and usage guidelines. Modal shift is an important function in managing mobile traffic in densely populated areas and locations with seasonal traffic peaks.

The development of HetNets is closely linked to the development of mobile and fixed networks. Operator WLAN hotspots are provided in city centers, business districts and certain event locations such as stadiums, convention centers, transport hubs and airports, for example. Not only do these hotspots encourage the use of an operator’s Wi-Fi mobile services, but they also help move cellular traffic to the operator’s landlines, reducing the stress on cellular base stations from the increased traffic.

In other HetNet scenarios, operators operate 2G and 3G legacy networks alongside LTE and 5G. Such coexistence can be attributed to factors such as sunk costs, old plan contracts, capacity optimization, monetization strategies of the operators and the longer time delay until a new network is available nationwide. Offloading from 5G to LTE or LTE to 3G is common in these scenarios as user traffic shifts from congested to available networks.

In densely built-up areas, such as city centers with multi-story skyscrapers, multi-story highways, and underground transportation, concrete structures and space constraints limit macrocell coverage. In these areas, operators are shifting traffic to either small cells or DAS, both of which can be used more flexibly on buildings and common urban structures such as bus stops and lamp posts. Discharging takes place between macro cells, small cells and DAS in this scenario.

Make space for better decisions when unloading

The decisions to shift traffic between these different access networks are underpinned by real-time network transparency and application awareness through tools such as R & S®PACE 2. R & S®PACE 2, deep packet inspection (DPI) software, offers a detailed overview of the data traffic in a cellular network – whether on the access, transport or core layer – via traffic classification and the extraction of metadata. This enables operators to decide how, what and when to unload.

The R & S®PACE 2 traffic classification enables applications to be identified. Latency-sensitive applications such as autonomous driving can be affected by inefficient switchovers under blanket offloading rules in which traffic is shifted back and forth between available networks. At the same time, bandwidth-intensive, but not latency-sensitive applications, such as video streaming, video calls or file downloads, can be outsourced at the first opportunity, freeing up capacity in the cellular network. Both scenarios require intelligent offloading of mobile traffic, with offloading based on the type and specific attributes of an application, such as a video call in a chat application.

Intelligent offloading also extends to other access nodes such as small cells. In a demonstration by research engineers in the Intel Lab, the integration of R & S®PACE 2 in their “smart pipe” server used in small cells made it possible to identify specific applications and traffic types in real time. This allows traffic to be shifted to other access points such as available macro cells and operator Wi-Fi hotspots or vice versa, making it possible to prioritize selected applications such as Skype and YouTube and better manage traffic in crowded business districts and residential high-rise clusters.

R & S®PACE 2 also offers metadata extraction, which enables mobile network operators to determine the status of the network at each access node. This includes information on bandwidth consumption, speeds, latency and jitter. Access nodes with high congestion, sudden spikes in traffic, or physical problems such as power outages or hardware failures can be identified before degradation in service becomes apparent. Routing traffic to alternate nodes using information from DPI enables operators to balance traffic spikes and dips and maximize network capacity by using all of the existing infrastructure while ensuring a consistent quality of service for the user.

Class segregation

Interestingly, R & S®PACE 2 goes one step further by making decisions about mobile data offloading easier by identifying the generic devices used. By leveraging this information and combining it with user and device information provided in protocols like GTP, operators can design their traffic management policies to prioritize connections for users on Premium plans and provide higher quality of service (QoS) by using on routes at the highest speeds and by optimizing content for the devices used. Likewise, users who have subscribed to cell phone tariffs with free WiFi access are correctly and promptly identified and forwarded to WiFi hotspots as soon as they are available, which saves subscription fees for data usage.

Not all traffic is good traffic

The R & S®PACE 2 combines traffic classification and anomaly detection support, providing real-time inputs that can be used by network security functions such as firewalls and intrusion detection and prevention systems to manage and prevent cyberattacks on the network. R & S®PACE 2 gets to the bottom of such attacks by identifying the generic devices used and the intensity and frequency of such attacks. These findings are crucial for outsourcing decisions. They enable mobile network operators to correctly and promptly determine whether the data traffic is suspicious or abnormal and whether this data traffic should be moved to a less congested network, routed through additional firewalls or blocked entirely.

R & S®PACE 2, with its traffic analysis capabilities and support for anomaly detection, plays another key role in traffic offloading. It provides real-time traffic insights that operators can use to automate the authentication of users connecting to their Wi-Fi hotspots. More specifically, the information from DPI helps operators protect Wi-Fi access points from malicious traffic, Denial of Service (DoS) attacks, illegal tethering, and other fraudulent uses. With 5G offloading, DPI goes one step further by providing the information necessary to instantiate the correct Wi-Fi QoS slice. For example, a low-latency Wi-Fi slice must provide instant authentication and expedited processing to ensure end-to-end, consistent mobile performance.

In many ways, DPI perfects mobile data offloading. It provides the information necessary for intelligent decision making, increasing network efficiency and optimization. Long-term data provided by DPI can be customized into improved offloading guidelines, taking into account the frequency, intensity, and application usage trends at each access node. Not only does it pave the way for improved cellular network performance and coverage, it also enables barely noticeable and seamless offloading for any mobile user.

You might also like

Comments are closed.