Aricent finds ways to extend the life of your 4G investments through integration with cloud technology.
In 2014, for the first time, the number of mobile gadgets on the planet surpassed the number of people. Today, there are 8.5 billion mobile connections and 7.6 billion people. The number of gadgets is multiplying five times faster than the world’s population, producing mind-boggling amounts of data. Soon, the current infrastructure provisioning could burst at the seams as the overall data capacity cannot keep up with the amount of data volume created. It is imperative to start planning and implementing a new paradigm for connecting all these devices and managing the data they produce.
Essentially, this ever-expanding hunger for data results in pain points for all the stakeholders involved including:
- End Users get frustrated when their calls don’t go through or get dropped, web browser freezes, 3G/4G data speeds are low and inconsistent. It seems Murphy’s Law is commonplace.
- Operators cannot realize the full potential of their deployed hardware or derive maximum average revenue per unit(ARPU). All heuristics are deployed to achieve maximum optimization, but the situation keeps changing and network planning can’t be changed every week. To meet the network requirements, operators have to physically install new infrastructure which takes a lot of time. Installing a new tower can even take six to 12 months or more for approvals and physical setup, causing loss of customers and revenue.
Utilization of the network infrastructure varies according to the day of the week and time of day. In an area populated predominantly by businesses, infrastructure is utilized to the hilt on weekdays and virtually unused on weekends. This usage pattern is reversed in residential areas, where usage is high in the evenings and on weekends. In malls and shopping complexes, usage is low during weekdays and high on weekends and holidays.
Networks are plagued by dropped calls and poor quality connections. Increasing pressure from regulatory authorities is forcing operators to control dropped calls, which results in operators building more towers. To manage this situation, operators use hardware resources sub-optimally, by over-provisioning the hardware and software resources. Operators try and execute long-term network planning, but it is hard to predict and plan accurately. Investments in infrastructure can’t keep pace with the rate of growth of mobile devices, customer usage and the uneven usage patterns in different locations and at different times of the day.
To make the network future-proof—at least to some extent—cloud technologies are applied to the access side—evolved packet core (EPC)—to cater to traffic increases and sudden bursts. However, for the elements in E-UTRAN (which is the combination of the user equipment, eNodeB, and the network), the only option today is to increase the number of towers or use more spectrum, which is in limited supply, expensive and hence economically unsustainable.
Emerging Best Practice
Despite these challenges, it is possible to reduce congestion by optimizing the network infrastructure.
Consider the 4G Network. An eNB helps a UE (user element/mobile device) to access the radio channel and create a link between the UE and the LTE core network, also called the EPC. The EPC architecture consists of multiple modules, such as MME, PCRF, S-GW, and P-GW, which helps separate the control plane (CP) from the data plane (U-Plane).
In the near future, the ever-increasing number of mobile devices and applications will put an increasing amount of strain on already over-strained mobile networks thus aggregating the problem of network congestion. There are some solutions to help reduce, if not totally resolve, these challenges:
- Additional spectrum allocation. This is a very expensive option but would be used to cater to the growing needs for access, especially in very high population areas where mobile device density is high.
- Making 4G calls over a WiFi network - this will help a mobile user make a call using Wi-Fi network and route it to the destination mobile or landline network. This reduces the load on the 4G network, hence freeing up spectrum for additional purposes.
- Additional eNB. More eNBs would help but installing them takes time and as tower’s range is decreasing from micro to pico distances, density increases with no additional spectrum allocation. Moreover, the increase in the number of eNBs would also result in more power consumption and higher radiation exposure which could be detrimental.
- Network and frequency re-configuration of eNodeB. This can improve coverage improve coverage, but it is complex to reconfigure and maintain.
- Evolved Packet Core (EPC) network virtualization. A framework where the functions required to converge data and voice on 4G LTE networks are moved from dedicated hardware to software that operates on low-cost carrier grade COTS hardware.
- Internet congestion reduction. This can be accomplished using a content delivery network (CDN)and fog computing to move delay sensitive or high data consuming applications from the internet and cloud to EPCs.
We propose the following recommendations for balancing and optimizing eNodeB resources, which can be done independently and together as well.
Recommendation A: Have a separate spectrum allocation manager (SAM). The SAM would perform real-time capacity adjustments, spectrum allocation and movement across eNodeB. The SAM is deployed in the EPC and reallocates spectrum based on network conditions in a region consisting of a set of eNodeBs. The SAM replaces the manual configuration done by the network management systems today and can evolve to an automatic system that leverages machine learning and analytics, which can reconfigure and tune the system on need basis.
Aricent has developed EPC intellectual property for use in the Open eNodeB solution. Changes can be made in the EPC to demonstrate the proof of concept.
3GPP Rel 9 and 10 have introduced self-optimization functions like Mobility Load Balancing (MLB) in which cells with congestion can handover load to other cells, which have spare resources. MLB is also used between different Radio technologies in which case the RAN Information protocol (RIM) is used to transfer information between cores of various base stations which may be running 3G or 4G.
Each eNodeB detects congestion on U-Plane and accordingly does a partial UE handoff i.e forwards the UE U-Plane traffic to less loaded eNodeBs, instead of doing a completely handoff to the other eNB. This technique would be beneficial when the IP based U-Plane is congested for particular eNBs.
We propose that Option 2 be used as it is economically viable for operators to increase cell density with lesser Capex / Opex expenditures
We believe that marrying cloud technologies with the current 4G network infrastructure will help prolong the lifespan of 4G, deliver expanded service capabilities, scalability and improved management of available resources, thus giving operators more time to amortize 4G investments, while 5G beckons in the future.
About the Author
Gurjit S. Butalia and Ravi Dubey
Gurjit is AVP of Technology at Aricent with 21 years of experience in system software, embedded, web and cloud technologies. He specializes in architecting applications for Cloud and is an AWS certified Architect.
Ravi is Director of Engineering at Aricent with 21 years of experience in wireless domain, system software in Linux and other embedded RTOSs. He specializes in design, development of platform software which is in use by various 3G, 4G, 5G network elements of leading Telecom vendors. He is currently working on 4G platform and Microwave radio based Backhaul system development.