• Venky

Deep Packet Inspection - Still a 5G requirement ?

Updated: Feb 13

Venkatesh (Venky) Aravamudhan

Explosion of data traffic is evident in today's world with people working from home, streaming while travelling and so on. Radio enhancements always enhanced the user experience wherein users have moved from sending a small multi media message in the GPRS/EDGE era to real media media access and High Definition (HD) content in the LTE/5G NSA.

Initial use case of Deep Packet Inspection (DPI) was utilized for differentiating traffic thereby services to differentially charge subscribers, control QoS and throttle when needed. Creation of packs such as Social networking to provide additional streams for monetization. DPI use cases enhanced through the well known limitation of TCP performance in low speed connection and a mechanism to enhance the performance. Utilization of well known TCP algorithms to enhance the TCP state mechanism to provide latency and throughput benefits was implemented on the user plane function provided solution to alleviate the damage. Use case pertaining to subscriber level Policy, QoS, Proxy function...etc. created core networks to spawn more data processing capabilities than the control plane. This obviously lead to 3GPP creating Control Plane User Plane segregation (CUPS) in the core network in order to help operator scale data plane nodes separately than the control plane entities.

Day without Internet is like a Year with no Rain (Internet Quote)

Most smartphone manufacturers are releasing at least two new models every year indicates that most users estimate data as a minimum requirement across the globe. Currently with emergence of 5G radio, with projected support for 1Gbps worth of throughput per UE, use cases pertaining to differential treatment based on specific service, monetization and poor radio quality may soon disappear (with considerable bandwidth on radio). However, it has to be noted that government agencies across the world mandate safe internet for children, content filtering solution, which are still valid and necessary requirements which do not go away with the 5G Radio (or any access).

It is inevitable that mobile operators still require DPI and actions based on the necessary compliances that needs to be adhered. However, performing DPI for all user data is not proper utilization of the compute resources. Internet communities are implementing security mechanisms to hide traffic through encryptions such as TLS schemes thereby lesser traffic may be analyzed directly, DPI engines can work only through patterns at best.

It is envisaged that 5G core deployment model would include mechanisms to differentiate user traffic based on 3GPP defined parameters such as slice/DPN(DNN) . Operator can group subscribers and services that demand compliance with internet security in turn DPI and analysis. Irrespective of the radio access technology, mobile networks would include traffic that needs to be analyzed, inspected and will need necessary actions hence DPI requirements remain. Having said that DPI is required, does it need to be performed for all user plane traffic remains a big question. Operators deployment model should consider these factor in order to have a highly efficient network .

110 views0 comments

Recent Posts

See All

If you haven't heard about Apple Airtag, maybe you missed an exciting upcoming tech. Airtag connects via Bluetooth with other Apple devices. What's interesting is that Airtags can connect to any Apple

Edging towards the first full-fledged commercial 5G deployment, this seems like a good time to look at the last 3 years of 5G hype. Let's start with some prophecies and ideas that Catenna has been ad

In the last two years, 3GPP has been tirelessly working to standardize every aspect of 5G core network. With Release 16 now frozen, it seems like a lot has been achieved. With the CSPs increasingly wo