Pond: Cxl-based memory pooling systems for cloud platforms

H Li, DS Berger, L Hsu, D Ernst, P Zardoshti… - Proceedings of the 28th …, 2023 - dl.acm.org
Public cloud providers seek to meet stringent performance requirements and low hardware
cost. A key driver of performance and cost is main memory. Memory pooling promises to …

[HTML][HTML] {SONIC}: Application-aware data passing for chained serverless applications

A Mahgoub, L Wang, K Shankar, Y Zhang… - 2021 USENIX Annual …, 2021 - s.usenix.org
The conference papers and full proceedings are available to registered attendees now and
will be available to everyone beginning Wednesday, July 14, 2021. Paper abstracts and …

From cloud to edge: a first look at public edge platforms

M Xu, Z Fu, X Ma, L Zhang, Y Li, F Qian… - Proceedings of the 21st …, 2021 - dl.acm.org
Public edge platforms have drawn increasing attention from both academia and industry. In
this study, we perform a first-of-its-kind measurement study on a leading public edge …

Providing {SLOs} for {Resource-Harvesting}{VMs} in cloud platforms

P Ambati, Í Goiri, F Frujeri, A Gun, K Wang… - … USENIX Symposium on …, 2020 - usenix.org
Cloud providers rent the resources they do not allocate as evictable virtual machines (VMs),
like spot instances. In this paper, we first characterize the unallocated resources in Microsoft …

Assess and summarize: Improve outage understanding with large language models

P Jin, S Zhang, M Ma, H Li, Y Kang, L Li, Y Liu… - Proceedings of the 31st …, 2023 - dl.acm.org
Cloud systems have become increasingly popular in recent years due to their flexibility and
scalability. Each time cloud computing applications and services hosted on the cloud are …

Cost-efficient overclocking in immersion-cooled datacenters

M Jalili, I Manousakis, Í Goiri, PA Misra… - 2021 ACM/IEEE 48th …, 2021 - ieeexplore.ieee.org
Cloud providers typically use air-based solutions for cooling servers in datacenters.
However, increasing transistor counts and the end of Dennard scaling will result in chips …

Beware of Fragmentation: Scheduling {GPU-Sharing} Workloads with Fragmentation Gradient Descent

Q Weng, L Yang, Y Yu, W Wang, X Tang… - 2023 USENIX Annual …, 2023 - usenix.org
Large tech companies are piling up a massive number of GPUs in their server fleets to run
diverse machine learning (ML) workloads. However, these expensive devices often suffer …

{SelfTune}: Tuning Cluster Managers

A Karthikeyan, N Natarajan, G Somashekar… - … USENIX Symposium on …, 2023 - usenix.org
Large-scale cloud providers rely on cluster managers for container allocation and load
balancing (eg, Kubernetes), VM provisioning (eg, Protean), and other management tasks …

Learning to schedule multi-NUMA virtual machines via reinforcement learning

J Sheng, Y Hu, W Zhou, L Zhu, B Jin, J Wang… - Pattern Recognition, 2022 - Elsevier
With the rapid development of cloud computing, the importance of dynamic virtual machine
scheduling is increasing. Existing works formulate the VM scheduling as a bin-packing …

Growable Genetic Algorithm with Heuristic-based Local Search for multi-dimensional resources scheduling of cloud computing

G Zhou, WH Tian, R Buyya, K Wu - Applied Soft Computing, 2023 - Elsevier
Abstract Multi-Dimensional Resources Scheduling Problem (MDRSP, usually a multi-
objective optimization problem) has attracted focus in the management of large-scale cloud …