This paper presents a platform for measuring, evaluating and modeling the energy efficiency aspects of an O-RAN 5G wireless network. We describe our open-source based O-RAN testbed which includes both bare-metal and Kubernetes network functions, in addition to physical network components. We focus on measuring power consumption of servers and workloads of cloudified and virtualized network functions. We show that a combination of different power measurements can be used successfully to achieve the accuracy and granularity required for energy efficiency measurement, evaluation and modeling.

Modern communication networks are large, dynamic, complex, and increasingly use virtualized network infrastructure. To deploy, maintain, and troubleshoot such networks, it is essential to understand how network elements - such as servers, switches, virtual machines, and virtual network functions - are connected to one another, and to be able to discover communication paths between them. For network maintenance applications such as troubleshooting and service quality management, it is also essential to understand how connections change over time, and be able to pose time-travel queries to retrieve information about past network states. With the industry-wide move to Software Defined Networks and Virtualized Network Functions (VNFs) [26][24], maintaining these inventory and topology databases becomes a critical issue.

Modern communication networks are large, dynamic, and complex. To deploy, maintain, and troubleshoot such networks, it is essential to understand how network elements such as servers, switches, virtual machines, and virtual network functions are connected to one another, and to be able to discover communication paths between them. For network maintenance applications such as troubleshooting and service quality management it is also essential to understand how connections change over time, and be able to pose time-travel queries to retrieve information about past network states. With the industry-wide move to SDNs and virtualized network functions [13], maintaining these inventory databases becomes a critical issue.

LTE-Advanced (LTE-A) macro-cell deployments are being enhanced with small cells, i.e., low-power base stations, to increase the network coverage and capacity. However, simultaneous co-channel transmissions from macro and small cells cause increased inter-cell interference and under-utilize the spectrum resources at the small cells. The following LTE-A design techniques are used to improve system performance in such deployments: (i) Carrier Aggregation (CA) to increase capacity by using additional carrier bandwidth; (ii) enhanced Inter-Cell Interference Coordination (elCIC), that includes (a) Cell Selection Biasing (CSB) to increase small cell spectrum utilization via cell range expansion; and (b) blanking data transmission on the macro cells for a certain duration of time to increase cell-edge user throughput Our objective is to maximize the CSB of the small cell, subject to user QoS constraints and blanking support from the macro cell. Towards this end, we develop an analytical model that captures the inter-dependency between elCIC techniques. We observe that, not accounting for the complex inter-dependencies between these techniques leads to a degraded network performance. We propose a framework that jointly optimizes elCIC and the assignment of multiple component carriers in an LTE-A deployment for increasing spectrum utilization at the small cells with appropriate blanking support from the macro cells. Our simulation results show that our approach increases the small cell spectrum utilization and aggregate cell-edge throughput by as much as 200%.

HTTP-based adaptive protocols dominate today's video streaming over the Internet, and operate using multiple quality levels that video players request one segment at a time. Despite their popularity, studies have shown that performance of video streams still suffers from stalls, quality switches and startup delay. In wireless networks, it is well-known that high variability in network bandwidth affects video streaming. MultiPath TCP (MPTCP) is an emerging paradigm that could offer significant benefits to video streaming by combining bandwidth on multiple network interfaces, in particular for mobile devices that typically support both WiFi and cellular networks. In this paper, we explore whether MPTCP always benefits mobile video streaming. Our experimental study on video streaming using two wireless interfaces yields mixed results. While beneficial to user experience under ample and stable bandwidth, MPTCP may not offer any advantage under some network conditions. We find that when additional bandwidth on the secondary path is not sufficient to sustain an upgrade in video quality, it is generally better not to use MPTCP. We also identify that MPTCP can harm user experience when an unstable secondary path is added to the stable primary path.

Proportional Fair (PF) scheduling algorithms are the de facto standard in cellular networks. They exploit the users' channel state diversity (induced by fast-fading) and are optimal for stationary channel state distributions and an infinite time-horizon. However, mobile users experience a nonstationary channel, due to slow-fading (on the order of seconds), and are associated with base stations for short periods. Hence, we develop the Predictive Finite-horizon PF Scheduling ((PF)2S) Framework that exploits mobility. We present extensive channel measurement results from a 3G network and characterize mobility-induced channel state trends. We show that a user's channel state is highly reproducible and leverage that to develop a data rate prediction mechanism. We then present a few channel allocation estimation algorithms that exploit the prediction mechanism. Our trace-based simulations consider instances of the (PF)2S Framework composed of combinations of prediction and channel allocation estimation algorithms. They indicate that the framework can increase the throughput by 15%-55% compared to traditional PF schedulers, while improving fairness.

Planned upgrades in cellular networks occur every day, may often need to be performed on weekdays, and can potentially degrade service for customers. In this paper, we explore the problem of tuning network configurations in order to mitigate any potential impact due to a planned upgrade which takes the base station off-air. The objective is to recover the loss in service performance or coverage which would have occurred without any modifications. To our knowledge, impact mitigation for planned base station downtimes has not been explored before in the literature. The primary contribution of this work is a proactive approach based on a predictive model that uses operational data of user density distributions and path loss (rather than idealized analytical models of these) to quickly estimate the best power and tilt configuration of neighboring base stations that enables high recovery. A secondary contribution is an approach to minimize synchronized handovers. These ideas, embodied in a capability called Magus, enables us to recover up to 76% of the potential performance loss due to planned upgrades in some cases for a large US mobile network, and this recovery varies as a function of base station density. Moreover, Magus is able to reduce synchronized handovers by a factor of 8.

LTE network service reliability is highly dependent on the wireless coverage that is provided by cell towers (eNB). Therefore, the network operator's response to outage scenarios needs to be fast and efficient, in order to minimize any degradation in the Quality of Service (QoS). In this paper, we propose an outage mitigation framework for LTE-Advanced (LTE-A) wireless networks. Our framework exploits the inherent design features of LTE-A; it performs a dual optimization of the transmission power and beamforming weight parameters at each neighbor cell sector of the outage eNBs, while taking into account both the channel characteristics and residual eNB resources, after serving its current traffic load. Assuming statistical Channel State Information about the users at the eNBs, we show that this problem is theoretically NP-hard; thus we relax it as a convex optimization problem and solve for the optimal points using an iterative algorithm. Contrary to previously-proposed power control studies, our framework is specifically designed to alleviate the effects of sudden LTE-A eNB outages, where a large number of mobile users need to be efficiently offloaded to nearby towers. We present the detailed analytical design of our framework, and we assess its efficacy via extensive NS-3 simulations on an LTE-A topology. Our simulations demonstrate that our framework provides adequate coverage and QoS across all examined outage scenarios.

Cellular providers are rapidly deploying multiple technologies like cell biasing, carrier aggregation, co-ordinated interference control/scheduling to improve capacity and coverage. In this paper, we explore a complementary transport layer approach based on multipath TCP that can concurrently use multiple interfaces to boost throughput of users with poor coverage and improve fairness. Multipath TCP has been recently standardized by IETF and requires no modifications to applications. It has been shown to improve fairness and throughput in wire line environments and individual user throughputs in wireless networks. However, in a wireless multi-user environment, it is not clear that it is always beneficial, as we show in this paper. Therefore, we examine if it is indeed beneficial for a service provider to judiciously decide whether to enable multiple cellular interfaces on a smart phone based on a global centralized view of its network. Alternatively, should a device decide independently based only on a local view? To quantify the network wide impact in a system where users have multiple cellular interfaces, we have developed centralized and distributed heuristic algorithms to evaluate this, particularly in the context of fairness across all the users. Our simulations and numerical models show that there are potential gains in fairness (15-30%) to be realized by judiciously enabling multipath connections at the cell edge. These gains diminish as the number of users in a cell increases or users behave greedily. We also quantify the delicate balance between throughput and fairness. Our analysis provides an intuition on which user(s) in a cellular network stand to benefit the most by enabling multiple interfaces. We also discuss LTE protocol mechanisms to enforce associations of specific interfaces to specific cells.

Proportional Fair (PF) scheduling algorithms are the de-facto standard in cellular networks. They exploit the users' channel state diversity (induced by fast-fading), and are optimal for stationary channel state distributions and an infinite time-horizon. However, mobile users experience a non-stationary channel, due to slow-fading (on the order of seconds), and are associated with basestations for short periods. Hence, we develop the Predictive Finite-horizon PF Scheduling ((PF)2S) Framework that exploits mobility. We present extensive channel measurement results from a 3G network and characterize mobility-induced channel state trends. We show that a user's channel state is highly reproducible and leverage that to develop a data rate prediction mechanism. We then present a few channel allocation estimation algorithms that rely on the prediction mechanism. Our trace-based simulations consider instances of the PF2S Framework composed of combinations of prediction and channel allocation estimation algorithms. They indicate that the framework can increase the throughput by 15%–55% compared to traditional PF schedulers, while improving fairness.

Full duplex communication requires nodes to cancel their own signal which appears as an interference at their receive antennas. Recent work has experimentally demonstrated the feasibility of full duplex communications using software radios. In this paper, we address capacity comparisons when the total amount of analog radio hardware is bounded. Under this constraint, it is not immediately clear if one should use these radios to perform full-duplex self-interference cancellation or use the radios to give additional MIMO multiplexing advantage. We find that repurposing radios for cancellation, instead of using all of them for half-duplex over-the-air transmission, can be beneficial since the resulting full-duplex system performs better in some practical SNR regimes and almost always outperforms half duplex in symmetric degrees-of-freedom (large SNR regime).

Server side measurements from several Wi-Fi hot-spots deployed in a nationwide network over different types of venues from small coffee shops to large enterprises are used to highlight differences in traffic volumes and patterns. We develop a common modeling framework for the number of simultaneously present customers. Our approach has many novel elements: (a) We combine statistical clustering with Poisson regression from Generalized Linear Models to fit a non-stationary Poisson process to the arrival counts and demonstrate its remarkable accuracy; (b) We model the heavy tailed distribution of connection durations through fitting a Phase Type distribution to its logarithm so that not only the tail but also the overall distribution is well matched; (c) We obtain the distribution of the number of simultaneously present customers from an Mt/G/∞ queuing model using a novel regenerative argument that is transparent and avoids the customarily made assumption of the queue starting empty at an infinite past; (d) Most importantly, we validate our models by comparison of their predictions and confidence intervals against test data that is not used in fitting the models.

The end to end system data performance over a 3G cellular network depends on many factors such as the number of users, interference, multipath propagation, radio resource management techniques as well as the interaction between these mechanisms and the transport protocol's flow and congestion mechanisms. Using controlled experiments in a public cell site, we investigate the interaction between TCP and the 3G UMTS/HSPA network's resource allocation, and its effect on fairness in the throughput achieved across multiple (up to 26) TCP flows in a loaded cell sector. Our field measurement results indicate that TCP fairness fluctuates significantly when the air interface (radio link) is the bottleneck. We also observe that TCP fairness is substantially better when the backhaul link (a fixed wired link) is the bottleneck, instead of the air interface. We speculate that the fairness of TCP flows is adversely impacted by the mismatch between the resource allocation mechanisms of TCP's flow and congestion control and that of the Radio Access Network (RAN).

In a dense WLAN environment, the signal coverage area of each access point (AP) typically has significant overlap with that of the neighboring APs. This is a problem if there are limited frequency channels. This paper presents an algorithm that can improve per-user throughput significantly, particularly for nonuniform traffic conditions. It is based on a cellular neural network model. Like a cellular neuron changing its state, based on the information of its neighboring neurons, every AP determines the best channel it should use in the next time slot, based solely on the traffic load of its neighboring APs and the channels used by them in the current time slot, but it actually switches to that channel with some fixed probability less than one. All APs in the network perform the above operation simultaneously. Computer simulations show that (1) given any traffic load distribution and any initial channel allocation, the algorithm converges to an equilibrium state in a short time, in which the overall throughput of the network is significantly improved; and (2) there exists an optimal switching probability that can minimize the time for the algorithm to reach the equilibrium state. The proposed technique has significant practical value due to its simplicity and effectiveness.

Wireless data networks will have a mix of user workloads since there will be a variety of wireless data terminal devices. Using computer simulation, we study the performance of an EDGE-based packet wireless network with various combinations of Web and dataphone users. We provide an understanding of the performance of a shared channel with mixed user workloads, and show that the number of users that can be supported can be determined using simple rules based on average traffic statistics.

Using analysis and simulation, we study the performance of Web-browsing and interactive data applications as perceived by the users of shared packet access networks. We use end-to-end page delays to characterise the user experience, and introduce a measure called equivalent circuit rate (ECR). The ECR for a user of a shared packet access network is the dedicated access circuit bit rate required to achieve the same user experience. Our analysis uses a simple closed queueing network model. The theoretical results agree well with simulations of TCP-based Web browsing workloads sharing an HFC cable data channel. We show that the mean page delay, ECR, and the number of users that can be supported are described using simple parameters that scale very well for different user and network rates. Our work also provides an effective comparison of shared and dedicated access at the application level.

This paper evaluates the use of packet scheduling to provide multiple service classes for bursty data traffic. In particular, we study the performance of the weighted fair queueing algorithm by simulating various scenarios in a packet cellular wireless network similar to an EDGE (Enhanced data rate for GSM evolution) system. The results show that, due to the burstiness of data traffic, traditional work-conserving packet scheduling algorithms such as weighted fair queueing cannot achieve the desired performance unless the system is heavily loaded. Similar results hold even when using priority queueing, which is an extreme form of weighted fair queueing. We also find that best effort traffic should be managed carefully to avoid inefficient usage of the network, especially in a low capacity cellular environment.

We describe an ongoing broadband fixed two-way wireless field experiment conducted by AT&T Laboratories Research, New Jersey, USA. Our experiment, which is one of the first two-way broadband fixed wireless systems anywhere, offers an end-to-end broadband packet access service, with telecommuting being the primary application for our employee users. It operates in the 2.6 GHz MMDS (Multichannel Multipoint Distribution Service) spectrum, and is based on cable modem technology. The lengthy period of operation has allowed us to make observations about user behavior, weather-related channel impairments, and equipment performance. We have identified several issues in the design of cable modem based systems which need to be tailored to the wireless environment. We have measured a significant path loss effect arising from a combination of rain and foliage. We have developed a Web based network monitoring/management tool that greatly enhances the ability to manage, diagnose, and optimize the system.

We consider use of fixed broadband wireless networks to provide packet services for telecommuting and Internet access. Each cell is divided into multiple sectors, each of them served by a sector antenna colocated with the base station (BS), and user terminals also use directional antennas mounted on the rooftops of homes or small offices and pointed to their respective BS antennas. To support a target data rate of 10 Mb/s, a bandwidth of several MHz is required. Since radio spectrum is expensive, the bandwidth needs to be reused very aggressively. Thus, efficient strategies for frequency reuse and managing cochannel interference are critically important. We propose several algorithms for dynamic radio-resource allocation in the fixed wireless networks. In particular, a method to be referred to as the staggered resource allocation (SRA) method uses a distributed scheduling algorithm to avoid major sources of interference while allowing concurrent packet transmission and meeting signal-to-interference objectives. The performance of the method is studied by analytic approximations and detailed simulation. Our results show that the combination of directional antennas plus the SRA method is highly effective in controlling cochannel interference. For reasonable system parameters, the SRA method delivers a throughput in excess of 30% per sector while permitting a given frequency band to be reused in every sector of every cell. It also provides satisfactory probability of successful packet transmission. In addition, a simple control mechanism can be applied in the method to improve performance for harsh radio environments.

Broadband analog transport facilities using fiber or fiber/coax cable can play a significant role in the evolution of the network infrastructure for personal communications services (PCSs). Low-power PCS systems require a dense grid of radio ports to provide connectivity to the telephone network. Analog transport has a number of important advantages over digital transmission facilities, including the flexibility to support a variety of air interface formats, shared infrastructure cost with other services such as video distribution, and centralized call processing allowing the use of low cost and simple radio ports. A simulcast technique can be used in such systems to permit low rates of handoff (no handoff within each simulcast area) and sharing of hardware resources among multiple radio ports. This paper provides a detailed model and a simulation analysis of the cochannel interference and noise performance as well as the resource sharing benefit of a simulcast PCS system. Several potential PCS air interfaces are considered, including time division multiple access (TDMA) and code division multiple access (CDMA) techniques. Our investigation shows that the impact of multiple antenna noise in a simulcast system is offset by the improved signal-to-interference (SIR) ratio brought about by distributed antennas. Even with distributed antennas, multiple antenna noise places a limit on the maximum number of radio ports that can be assigned to each simulcast group. This limit, however, is shown to have little impact on the achievable resource sharing benefit of simulcasting (i.e., grouping beyond this limit has diminishing returns). A saving of 40% to 60%, in terms of the required central hardware resources, is typical for both TDMA and CDMA systems in suburban environments.