Edge computing has become a very popular service that enables mobile devices to run complex tasks with the help of network-based computing resources. However, edge clouds are often resource-constrained, which makes resource allocation a challenging issue. In addition, edge cloud servers must make allocation decisions with only limited information available, since the arrival of future client tasks might be impossible to predict, and the states and behavior of neighboring servers might be obscured. We focus on a distributed resource allocation method in which servers operate independently and do not communicate with each other, but interact with clients (tasks) to make allocation decisions. We follow a two-round bidding approach to assign tasks to edge cloud servers, and servers are allowed to preempt previous tasks to allocate more useful ones. We evaluate the performance of our system using realistic simulations and real-world trace data from a high-performance computing cluster. Results show that our heuristic improves system-wide performance by 20-25% over previous work when accounting for the time taken by each approach. In this way, an ideal trade-off between performance and speed is achieved.

Through using edge computing services, mobile devices can run complex tasks with the help of network-based computing resources. However, servers in the edge cloud are not only constrained to limited resources, but also must make allocation decisions with only limited information available. The clients requesting computing resources may also have limited information about the servers available to them. We focus on a distributed resource allocation method in which servers operate independently and do not communicate with each other, but interact with clients to make allocation decisions for those clients’ tasks. We follow a two-round bidding approach to assign tasks to edge cloud servers. Servers may choose to preempt previous tasks to allocate more useful ones, and clients may choose to track the outcomes of their tasks to inform their future decisions. Results show that user learning improves system performance by 50-80% when servers are heterogeneous in pricing aggressiveness.

Competitive demand for network resources has only increased during the emergence of 5G next generation cellular technology. As video streaming accounts for an overwhelming percentage of this demand, the importance of considering the often-neglected Quality of Experience (QoE) metric is essential to ensure network resources are allocated in the most effective manner. Generalized network throughput metrics are insufficient in capturing the full human experience as increased data rates do not necessarily translate to improvements in user utility. Our study compares the efficacy of existing network allocation algorithms and proposes new approaches to 5G network resource allocation schemes using a more inclusive snapshot of user demand. We provide recommendations on which approach provides the highest QoE performance and suggestions for future network-side improvements. We further propose a QoE-driven network resource allocation (QENA) algorithm that shows a 20% improvement in overall average QoE across a large set of heterogeneous users.

Crowdsourced image processing has the potential to vastly impact response timeliness in various emergency situations. Because images can provide extremely important information regarding an event of interest (hits), sending the right images to an analyzer as soon as possible is of crucial importance. In this paper, we consider the problem of optimally assigning resources, both local (CPUs in phones) and remote (network-based GPUs) to mobile devices for processing images, ultimately sending those of interest to a centralized entity while also accounting for the energy consumption at the distributed nodes. To that end, we use the dual-path Network Utility Maximization (NUM) framework, coupled with a hit-ratio estimator and energy costs, to enable a distributed implementation of the system. We include analysis of different hit-ratio estimators using realistic trace data, first considering immediate and then delayed feedback. We address accuracy concerns when estimating the likelihood of future imagehitsand provide a window-based heuristic for scenarios when hit-ratio feedback is severely delayed. Our TCP-inspired window-method predicts both imagehitlikelihood and current wireless network congestion with great effectiveness. Results are validated using both synthetic simulations and real-life traces.

Crowdsourced image processing has the potential to vastly impact response timeliness in various emergency situations. Because images can provide extremely important information regarding an event of interest, sending the right images to an analyzer as soon as possible is of crucial importance. In this paper, we consider the problem of optimally assigning resources, both local (CPUs in phones) and remote (network-based GPUs) to mobile devices for processing images, ultimately sending those of interest to a centralized entity while also accounting for the energy consumption. To that end, we use the Network Utility Maximization (NUM) framework, coupled with a hit-ratio estimator and energy costs, to enable a distributed implementation of the system. Our results are validated using both synthetic simulations and real-life traces.

We introduce a novel mathematical model that treats network security as a game between cyber attackers and network administrators. The model takes the form of a zero-sum repeated game where each sub-game corresponds to a possible state of the attacker. Our formulation views state as the set of compromised edges in a graph opposed to the more traditional node-based view. This provides a more expressive model since it allows the defender to anticipate the direction of attack. Both players move independently and in continuous time allowing for the possibility of one player moving several times before the other does. This model shows that defense-in-depth is not always a rational strategy for budget constrained network administrators. Furthermore, a defender can dissuade a rational attacker from attempting to attack a network if the defense budget is sufficiently high. This means that a network administrator does not need to make their system completely free of vulnerabilities, they only to ensure the penalties for being caught outweigh the potential rewards gained.