Pond: Cxl-based memory pooling systems for cloud platforms
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 …
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
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 …
will be available to everyone beginning Wednesday, July 14, 2021. Paper abstracts and …
From cloud to edge: a first look at public edge platforms
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 …
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
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 …
like spot instances. In this paper, we first characterize the unallocated resources in Microsoft …
Assess and summarize: Improve outage understanding with large language models
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 …
scalability. Each time cloud computing applications and services hosted on the cloud are …
Cost-efficient overclocking in immersion-cooled datacenters
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 …
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
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 …
diverse machine learning (ML) workloads. However, these expensive devices often suffer …
{SelfTune}: Tuning Cluster Managers
Large-scale cloud providers rely on cluster managers for container allocation and load
balancing (eg, Kubernetes), VM provisioning (eg, Protean), and other management tasks …
balancing (eg, Kubernetes), VM provisioning (eg, Protean), and other management tasks …
Learning to schedule multi-NUMA virtual machines via reinforcement learning
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 …
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
Abstract Multi-Dimensional Resources Scheduling Problem (MDRSP, usually a multi-
objective optimization problem) has attracted focus in the management of large-scale cloud …
objective optimization problem) has attracted focus in the management of large-scale cloud …