When Should The Network Be The Computer?

Ports, Dan R. K.; Nelson, Jacob

doi:10.1145/3317550.3321439

Cited by 67 publications

(42 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A pipelined architecture is well-suited for networking protocols, which are typically characterized by relatively simple operations (e.g., rewrite the packet header or select an output port), but it is a poor fit for stateful in-network applications [38] because of the several implementation constraints that such an architecture introduces as we detail next.…”

Section: Limitations Of Pisamentioning

confidence: 99%

“…They have also paved the way for in-network computing [11,38,48], a new class of applications, ranging from caching [22] and database query processing [26,47] to machine learning (ML) [33,40,52] and consensus [16,17,27], that take advantage of the ability to execute arbitrary code within the network core (as opposed to just at the edge), leveraging the switches' unique vantage point. However, while programmable switches have been crucial to enable this new paradigm, we argue that their current architecture is a poor fit for the emerging applications, introducing unnecessary development complexity and impacting performance.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Challenging the Stateless Quo of Programmable Switches

Gebara

Lerner

Yang

et al. 2020

Proceedings of the 19th ACM Workshop on Hot Topics in Networks

View full text Add to dashboard Cite

Programmable switches based on the Protocol Independent Switch Architecture (PISA) have greatly enhanced the flexibility of today's networks by allowing new packet protocols to be deployed without any hardware changes. They have also been instrumental in enabling a new computing paradigm in which parts of an application's logic run within the network core (in-network computing). The characteristics and requirements of in-network applications, however, are quite different from those of packet protocols for which programmable switches were originally designed. Packet protocols are typically stateless, while in-network applications require frequent operations on shared state maintained in the switch. This mismatch increases the developing complexity of in-network computing and hampers widespread adoption. In this paper, we describe the key obstacles to developing innetwork applications on PISA and propose rethinking the current switch architecture. Rather than changing the existing architecture, we propose augmenting it with a Stateful Data Plane (SDP). The SDP supports the requirements of stateful applications, while the conventional data plane (CDP) performs packet-protocol functions.

show abstract

Section: Limitations Of Pisamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Challenging the Stateless Quo of Programmable Switches

Gebara

Lerner

Yang

et al. 2020

Proceedings of the 19th ACM Workshop on Hot Topics in Networks

View full text Add to dashboard Cite

show abstract

“…Our work is enabled by the long line of prior work on active networking [60], network function virtualization (NFV) [16], and programmable switches and SmartNICs [10,31,37,53,58] that enable programmability in the network fast path. Recent work on virtualizing programmable network devices and accelerators, including HyPer4 [25], P4Visor [65], AmorphOS [32] and AvA [63] have made it easier to share these offloads among applications.…”

Section: Related Workmentioning

confidence: 99%

Bertha: Tunneling through the Network API

Narayan

Panda

Alizadeh

et al. 2020

Proceedings of the 19th ACM Workshop on Hot Topics in Networks

View full text Add to dashboard Cite

Network APIs such as UNIX sockets, DPDK, Netmap, etc. assume that networks provide only end-to-end connectivity. However, networks increasingly include smart NICs and programmable switches that can implement both network and application functions. Several recent works have shown the benefit of offloading application functionality to the network, but using these approaches requires changing not just the applications, but also network and system configuration. In this paper we propose Bertha, a network API that provides a uniform abstraction for offloads, aiming to simplify their use.

show abstract

“…Network-Accelerated Systems. Recent projects have utilized programmable switches to accelerate consensus protocols [17,20,21,22], implement in-network caching [23], DNN training and inferencing [24], and in-network aggregation operations [25]. R2P2 [26] is the closest of these efforts to our design.…”

Section: Related Workmentioning

confidence: 99%

Falcon

Kettaneh

Udayashankar

Abdel-hadi

et al. 2020

Proceedings of the 3rd P4 Workshop in Europe

View full text Add to dashboard Cite

We present Falcon, a novel scheduler design for large scale data analytics workloads. To improve the quality of the scheduling decisions, Falcon uses a single central scheduler. To scale the central scheduler to support large clusters, Falcon offloads the scheduling operation to a programmable switch. The core of the Falcon design is a novel pipeline-based scheduling logic that can schedule tasks at line-rate. Our prototype evaluation on a cluster with a Barefoot Tofino switch shows that the proposed approach can reduce scheduling overhead by 26 times and increase the scheduling throughput by 25 times compared to state-of-the-art centralized and decentralized schedulers. CCS Concepts • Networks → Network services → In-network processing; • Networks → Network services → Cloud computing;

show abstract

When Should The Network Be The Computer?

Cited by 67 publications

References 19 publications

Challenging the Stateless Quo of Programmable Switches

Challenging the Stateless Quo of Programmable Switches

Bertha: Tunneling through the Network API

Falcon

Contact Info

Product

Resources

About