On the Fairness-Efficiency Tradeoff for Packet Processing with Multiple Resources

Wang, Wei; Feng, Chen; Li, Baochun; Liang, Ben

doi:10.1145/2674005.2675010

Cited by 17 publications

(9 citation statements)

References 41 publications

(97 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Wang et al studied the trade-off between multi-resource fairness and resource utilization, both theoretically [67] and in storage systems [66]. However, like other multi-resource scheduling algorithms developed for data centers [23,28,38,65], the latter is not directly applicable to our setting because our system must make online scheduling decisions.…”

Section: Related Workmentioning

confidence: 99%

Sandpaper

Helt

Feng

Seshan

et al. 2019

Proceedings of the 4th ACM/IEEE Symposium on Edge Computing

View full text Add to dashboard Cite

Modern content delivery networks (CDNs) allow their customers (i.e., operators of web services) to customize the processing of requests by uploading and executing code at the edges of the CDN's network. To achieve scale, CDNs have forgone heavyweight virtualization techniques. Instead, all requests often execute within the same OS or even process. However, performance interference may arise when these requests have differing demands on multiple system resources. In this paper, we study the sources of performance interference based on workloads from real customers, identify the lack of multi-resource fairness as the culprit, and show that existing schedulers available in commodity OSs are insufficient to enforce fairness between customers.We then design Sandpaper, a new and practical multiresource request scheduler for mitigating performance interference in CDN edge environments. Sandpaper enforces fairness despite constraints, such as sitting within the application runtime and running atop the OS's underlying resource schedulers. By leveraging key insights about the differences between theoretical system models and real systems, Sandpaper bridges the trade-off between resource utilization and multi-resource fairness that plagues existing schedulers. We implement Sandpaper atop Varnish, an open-source CDN edge proxy, and show that it mitigates performance interference while maintaining high resource utilization and with little performance overhead.

show abstract

Section: Related Workmentioning

confidence: 99%

Sandpaper

Helt

Feng

Seshan

et al. 2019

Proceedings of the 4th ACM/IEEE Symposium on Edge Computing

View full text Add to dashboard Cite

show abstract

“…New multi-resource scheduling designs for MEC must account for this complexity. Furthermore, multi-resource scheduling necessarily leads to the tradeoff between fairness among jobs and system efficiency [25] [46], which does not exist in traditional single-resource scheduling with work conservation. This is further exacerbated by the heterogeneous computation and communication capabilities of local and remote equipment [18] [48].…”

Section: Multi-resource Schedulingmentioning

confidence: 99%

Mobile Edge Computing

Liang

2017

Key Technologies for 5G Wireless Systems

Self Cite

View full text Add to dashboard Cite

The ongoing development of the fifth-generation (5G) wireless technologies takes place in a unique landscape of recent advancement in information processing, marked by the emerging prevalence of cloud-based computing and smart mobile devices. These two technologies complement each other by design, with cloud servers providing the engine for computing and smart mobile devices naturally serving as human interface and untethered sensory inputs. Together, they are transforming a wide array of important applications such as telecommunication, industrial production, education, e-commerce, mobile healthcare, and environmental monitoring. We are entering a world where computation is ubiquitously accessible on local devices, global servers, and processors everywhere in between. Future wireless networks will provide communication infrastructure support to this ubiquitous computing paradigm, but at the same time they can also utilize the new-found computing power to drastically improve communication efficiency, expand service variety, shorten service delay, and reduce operation expenses. Previous generations of wireless networks are passive systems. Residing near the edge of the Internet, they serve only as communication access pathways for mobile devices to reach the Internet core and the public switched telephone network (PSTN). Improvements to these wireless networks have focused on the communication hardware and software, such as advanced electronics and signal processing in the transmitters and receivers. Even for 5G, substantial research effort has been devoted to densification techniques, such as small cells, device-todevice (D2D), and massive multi-input multi-output (MIMO). Successes of this communication-only wireless evolution reflect the classical view of an information age centered at information consumption through the Internet. Yet, in many emerging applications, communication and computation are no longer separated, but interactive and unified. For example, in an augmentedreality application, which might be displayed on smart eyeglasses , the user's mobile device continuously records its current view, computes its own location, and streams the combined information to the cloud server, while the cloud server performs pattern recognition and information retrieval and sends back to the mo

show abstract

“…Grandl et al [12] incorporated DRF (or one of its variants) into a multiresource packing scheduler that speeds up job completion without compromising fairness much. DRF has also been implemented as a packet scheduler in middleboxes and software routers [9], [27].…”

Section: Related Workmentioning

confidence: 99%

Multi-resource Fair Sharing for Datacenter Jobs with Placement Constraints

Wang

Liang

et al. 2016

SC16: International Conference for High Performance Computing, Networking, Storage and Analysis

Self Cite

View full text Add to dashboard Cite

Providing quality-of-service guarantees by means of fair sharing has never been more challenging in datacenters. Due to the heterogeneity of machine configurations, datacenter jobs frequently specify placement constraints, restricting them to run on a particular class of machines meeting specific hardware/software requirements. In addition, jobs have diverse demands across multiple resource types, and may saturate any of the CPU, memory, or storage resources. Despite the rich body of recent work on datacenter scheduling, it remains unclear how multi-resource fair sharing is defined and achieved for jobs with placement constraints. In this paper, we propose a new sharing policy called Task Share Fairness (TSF). With TSF, jobs are better off sharing the datacenter, and are better off reporting demands and constraints truthfully. We have prototyped TSF on Apache Mesos and confirmed its service guarantees in a 50-node EC2 cluster. Trace-driven simulations have further revealed that TSF speeds up 60% of tasks over existing fair schedulers.

show abstract

On the Fairness-Efficiency Tradeoff for Packet Processing with Multiple Resources

Cited by 17 publications

References 41 publications

Sandpaper

Sandpaper

Mobile Edge Computing

Multi-resource Fair Sharing for Datacenter Jobs with Placement Constraints

Contact Info

Product

Resources

About