Deep packet inspection (DPI) often alludes to network intelligence. DPI, being a tool that delivers identification and classification of IP traffic, is known for its capability of collating real-time data from multiple network points to deliver insights that make great inputs for network- and application-level decision making.
In the era of 5G, DPI’s role of providing network intelligence through fine-grained analytics is expected to be a major driver for monetization. For mobile operators, monetization is essentially a function of how product innovations, pricing strategies and service delivery meet the expectations of end users. With 5G, product innovations, pricing and delivery hinge on how well mobile operators leverage their new 5G architecture to deliver a wide range of disparate services (products) on a common network at the expected quality within the given resources. This is partly due to the fact that 5G, unlike its predecessors 3G, 4G and LTE Advanced, features multiple service categories defined by different speeds, latencies, mobility and coverage attributes. Broad classification of these attributes have resulted in the creation of three major classes, namely the evolved mobile broadband (eMBB), the ultra-reliable ultra-low latency communications (URLLC) and the massive machine type communications (mMTC).
An application that is delivered on 5G essentially uses the service class that best corresponds to its required speeds, latencies, mobility and coverage. For example, an application that streams live video from a stadium will require eMBB connectivity while a smart city application will require connectivity on the mMTC service layer to enable its data to be collected and delivered efficiently at the expected quality of experience (QoE). The ability of 5G to deliver speed- and latency-sensitive applications alongside a host of other generic, low priority applications is what defines 5G, and is what will drive its monetization.
In this regard, DPI delivers the critical capability of identifying each application in real-time. This allows mobile operators to route traffic along different network slices using 5G’s network slicing capability. Each network slice is logically partitioned from other slices to cater to specific traffic types. In terms of monetization, this very feature is what enables 5G service providers to offer their clients a slice of the network according to the clients’ requirements. A number of hospitals connected on 5G for example will receive a slice of URLLC which guarantees that remote surgery applications are delivered at millisecond latencies. All other traffic on the network is routed through other slices that best fit that traffic, according to classification information delivered by DPI. DPI hence plays a central role in driving service differentiation, and in enabling operators to earn premiums on slices that are built to cater for superior connectivity.
DPI real-time traffic detection functionality also enables Operators to execute innovative pricing models - not just pay per-use pricing types but also more dynamic pricing strategies. DPI’s ability to identify traffic by metadata extraction enables an application traffic to be routed by the application attribute, resulting in bandwidth heavy attributes such as video files to be delivered on eMBB while light content such as text to be delivered on standard routes. This allows an application to receive two sets of pricing, balancing costs and QoS for the end client, while maximising revenues for the mobile operator.
DPI’s metadata identification and classification capabilities drive monetization in many other ways. At this juncture, most of us are already sold on the idea of programmable networks for 5G, which refers to the ability of network providers to scale networks up and down, in terms of resources and network services, according to the current traffic needs. This level of scalability and flexibility is enabled by both software defined networking (SDN) which allows both centralized control and dynamic provisioning of network resources, and network functions virtualization (NFV) which enables proprietary hardware to now be deployed as virtualized network functions (VNFs) on commercial off-the-shelf (COTs) infrastructure. Traffic detection results and analytics provided by DPI are crucial for programmable networks as this information is used by SDN Controllers to optimize existing bandwidth, routes and network services such as firewalls and caching to match the traffic conditions and type of applications being delivered. This improves cost effectiveness of the overall network, resulting in bigger margins for the operators, and enhanced price competitiveness for their services.
DPI’s application- and network- awareness will also be powering multi-access edge computing (MEC) in 5G networks. Edge computing allows applications that are latency-sensitive to be processed at the edge, without having to navigate hundreds of network nodes to the network core. Applications such as AR/VR and face/vehicle recognition require edge computing for almost real-time delivery, and this calls for application awareness at the edge itself. Applications delivered at the network edge are expected to make up a huge portion of premium rated traffic on 5G, so with DPI, mobile operators will be able to capitalize on this demand which is expected to shoot up as 5G rollouts intensify across the globe.
In 5G, the pooling of computing and storage resources at the radio access network (RAN) via the virtualized RAN (vRAN) architecture is another key area where DPI comes into play. vRAN is developed to enable flexible allocation of these resources at the access network level. With 5G, vRAN plays the essential role of prioritizing the processing of critical applications, for example, those which are latency sensitive, whenever congestion is identified. DPI identifies not just the overall network conditions at the access network level, it also correlates the usage of bandwidth and network resources to establish dynamic prioritization of traffic according to the applications being delivered. In 5G, this capability becomes even more important as network densification is expected to see resources being swapped between many different access nodes clustered in the same locality. This capability will also determine smart handover decisions for bandwidth hogging applications that can be off-loaded for example, to operator wifi networks during peak traffic periods.
Past deployments of 2G, 3G and 4G has shown that the hype surrounding a new technology often starts wearing off once deployments begin as hype and all the promises are quickly replaced by the realities on the ground, especially those relating to monetization. Technologies such as DPI have been developed and perfected over time to provide mobile operators rolling out their brand new 5G networks a tool that helps them to address monetization in ways that allows them to deploy each inch of the network and every network resource therein to its best use. Armed with fine-grained, real-time insights derived through the analysis of every packet that crosses their network, mobile operators are able to push the monetization of their 5G networks to new levels with dynamic, responsive pricing and delivery models that will ensure not just increased profitability but enhanced QoE for all end customers.
For a more detailed read on now DPI enhances 5G and why it is a crucial enabler for monetization, download Rohde & Schwarz’s latest whitepaper ‘DPI and 5G: Network Visibility and Real-time Application Awareness’ for free.
This is a sponsored article.
