Concordia

Foukas, Xenofon; Radunović, Božidar

doi:10.1145/3452296.3472894

Cited by 38 publications

(24 citation statements)

References 64 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To exemplify, the processing time for decoding/encoding the users' uplink/downlink streams depends on the amount of this traffic which, most often, is subject to rapid fluctuations; see, e.g., [38] and our measurements in Fig. 1.…”

Section: Background and Motivationmentioning

confidence: 99%

“…Leveraging its high degree of data parallelization (intrinsic to its architecture) the GPU speeds up the TB processing, in contrast to the CPU where the processing is sequential (see [35] for more details). This delay variance is critical, as the vBS workloads are subject to stringent processing deadlines (1-3 ms) which, if violated, lead to data loss and energy waste [38,40]. Third, the energy cost of these computations depends on the PU technology and the data stream characteristics.…”

Section: Background and Motivationmentioning

confidence: 99%

“…These approaches have high overhead, e.g., require expensive matrix inversions, and provide optimality guarantees only under stationary conditions; therefore, they are typically employed for longer-term static resource control policies. Other practical solutions tailored to vRAN resource management include hybrid offline-trained and online-adapted vBS workload predictions [38], or regression models [40], so as to increase the utilization of the employed CPUs. These works do not provide optimality or fairness guarantees, and operate in real-time as opposed to the non-RT scale of our assignment policy; thus can be used concurrently.…”

Section: Resource Management In Vransmentioning

confidence: 99%

“…Here, we rely instead on the theory of online convex optimization (OCO) which: (i) does not require access to performance/cost functions or system/user-related parameters; and (ii) offers guarantees under a wide range of scenarios, including adversarial ones [92]. The robustness of OCO is particularly useful for vRANs which, due to the virtualized vBS functions, exhibit volatile performance and, importantly, high energy costs [18,26,38,40]. These experimental findings motivated the design of policies about the vBSs' transmission power, modulation and coding schemes, spectrum usage, and others, which aim to curb the vRANs energy costs [14,15,17].…”

Section: Resource Management In Vransmentioning

confidence: 99%

See 3 more Smart Citations

Fair Resource Allocation in Virtualized O-RAN Platforms

Aslan,

Iosifidis,

Ayala-Romero

et al. 2024

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

O-RAN systems and their deployment in virtualized general-purpose computing platforms (O-Cloud) constitute a paradigm shift expected to bring unprecedented performance gains. However, these architectures raise new implementation challenges and threaten to worsen the already-high energy consumption of mobile networks. This paper presents first a series of experiments which assess the O-Cloud's energy costs and their dependency on the servers' hardware, capacity and data traffic properties which, typically, change over time. Next, it proposes a compute policy for assigning the base station data loads to O-Cloud servers in an energy-efficient fashion; and a radio policy that determines at near-real-time the minimum transmission block size for each user so as to avoid unnecessary energy costs. The policies balance energy savings with performance, and ensure that both of them are dispersed fairly across the servers and users, respectively. To cater for the unknown and time-varying parameters affecting the policies, we develop a novel online learning framework with fairness guarantees that apply to the entire operation horizon of the system (long-term fairness). The policies are evaluated using trace-driven simulations and are fully implemented in an O-RAN compatible system where we measure the energy costs and throughput in realistic scenarios.

show abstract

Section: Background and Motivationmentioning

confidence: 99%

Section: Background and Motivationmentioning

confidence: 99%

Section: Resource Management In Vransmentioning

confidence: 99%

Section: Resource Management In Vransmentioning

confidence: 99%

See 2 more Smart Citations

Fair Resource Allocation in Virtualized O-RAN Platforms

Aslan,

Iosifidis,

Ayala-Romero

et al. 2024

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

show abstract

“…Other use cases for Libpreemptible include traffic shaping [61], scheduling in 5G [35], and real-time DNN serving [41]. The accuracy of these timed actions is crucial to the performance of these real-time application.…”

Section: B Other Use Cases and Future Workmentioning

confidence: 99%

Adaptive Task-Space Synchronization Control of SMMS Teleoperaiton Systems with Round-Robin Scheduling Protocol

Yin

et al. 2019

2019 Chinese Control Conference (CCC)

View full text Add to dashboard Cite

Modern cloud applications are prone to high tail latencies since their requests typically follow highly-dispersive distributions. Prior work has proposed both OS-and systemlevel solutions to reduce tail latencies for microsecond-scale workloads through better scheduling. Unfortunately, existing approaches like customized dataplane OSes, require significant OS changes, experience scalability limitations, or do not reach the full performance capabilities hardware offers. We propose LibPreemptible, a preemptive user-level threading library that is flexible, lightweight, and scalable. LibPreemptible is based on three key techniques: 1) a fast and lightweight hardware mechanism for delivery of timed interrupts, 2) a general-purpose user-level scheduling interface, and 3) an API for users to express adaptive scheduling policies tailored to the needs of their applications. Compared to the prior state-of-the-art scheduling system Shinjuku, our system achieves significant tail latency and throughput improvements for various workloads without the need to modify the kernel. We also demonstrate the flexibility of LibPreemptible across scheduling policies for real applications experiencing varying load levels and characteristics.

show abstract