What You Need to Know About (Smart) Network Interface Cards

Katsikas, Georgios P.; Barbette, Tom; Chiesa, Marco; Kostić, Dejan; Maguire, Gerald Q.

doi:10.1007/978-3-030-72582-2_19

Cited by 14 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In comparison with other accelerators (e.g., FPGA), direct processing of packets at NIC after their arrival, can omit the delay due to the transferring of packets from system memory to accelerator memory in order to be processed. However, the adoption of NIC for in-network computing still copes with challenges due to development process of applications as well as performance issues [46], [49]. In particular, to address the development abstraction in server application offloading, Floem [50] provides programming abstractions for the purpose of data placement, caching, and parallelism, and communication policies between program components across devices.…”

Section: Technologies Protocols Illustrative Use Cases and Criteriamentioning

confidence: 99%

A Survey on In-Network Computing: Programmable Data Plane and Technology Specific Applications

Kianpisheh

Taleb

2023

IEEE Commun. Surv. Tutorials

View full text Add to dashboard Cite

In comparison with cloud computing, edge computing offers processing at locations closer to end devices and reduces the user experienced latency. The new recent paradigm of innetwork computing employs programmable network elements to compute on the path and prior to traffic reaching the edge or cloud servers. It advances common edge/cloud server based computing through proposing line rate processing capabilities at closer locations to the end devices. This paper discusses use cases, enabler technologies and protocols for in-network computing. According to our study, considering programmable data plane as an enabler technology, potential in-network computing applications are in-network analytics, in-network caching, innetwork security, and in-network coordination. There are also technology specific applications of in-network computing in the scopes of cloud computing, edge computing, 5G/6G, and NFV. In this survey, the state of the art, in the framework of the proposed categorization, is reviewed. Furthermore, comparisons are provided in terms of a set of proposed criteria which assess the methods from the aspects of methodology, main results, as well as application-specific criteria. Finally, we discuss lessons learned and highlight some potential research directions.

show abstract

Section: Technologies Protocols Illustrative Use Cases and Criteriamentioning

confidence: 99%

A Survey on In-Network Computing: Programmable Data Plane and Technology Specific Applications

Kianpisheh

Taleb

2023

IEEE Commun. Surv. Tutorials

View full text Add to dashboard Cite

show abstract

“…Recent releases of NICs with connection tracking mechanisms [9] may open the path to implementations that exploit hardware offloading on commodity hardware, freeing the servers' CPU from the connection tracking duty [69].…”

Section: Future Workmentioning

confidence: 99%

High-speed Connection Tracking in Modern Servers

Girondi

Chiesa

Barbette

2021

2021 IEEE 22nd International Conference on High Performance Switching and Routing (HPSR)

Self Cite

View full text Add to dashboard Cite

The rise of commodity servers equipped with highspeed network interface cards poses increasing demands on the efficient implementation of connection tracking, i.e., the task of associating the connection identifier of an incoming packet to the state stored for that connection. In this work, we thoroughly investigate and compare the performance obtainable by different implementations of connection tracking using high-speed real traffic traces. Based on a load balancer use case, our results show that connection tracking is an expensive operation, achieving at most 24 Gbps on a single core. Core-sharding and lock-free hash tables emerge as the only suitable multi-thread approaches for enabling 100 Gbps packet processing. In contrast to recent beliefs, we observe that newly proposed techniques to "lazily" delete connection states are not more effective than properly tuned traditional deletion techniques based on timer wheels.Index Terms-Connection tracking, load balancer, high-speed networking, multi-core processing, hash table

show abstract

“…In addition to the large flow rule capacity of these NICs, all tables beyond table 0 have higher rule installation speeds. According to a recent study, the rule installation performance of the Mellanox NICs in table 1 can reach rates beyond half a million of rules per second using a single CPU core to inject NIC rules [47]. This performance can be further improved using multiple cores for rule injection [105].…”

Section: Impact Of Increasing the Number Of Traffic Classesmentioning

confidence: 99%

Metron

Katsikas

Barbette

Kostić

et al. 2020

ACM Trans. Comput. Syst.

Self Cite

View full text Add to dashboard Cite

Deployment of 100Gigabit Ethernet (GbE) links challenges the packet processing limits of commodity hardware used for Network Functions Virtualization (NFV). Moreover, realizing chained network functions (i.e., service chains) necessitates the use of multiple CPU cores, or even multiple servers, to process packets from such high speed links. Our system Metron jointly exploits the underlying network and commodity servers’ resources: ( i ) to offload part of the packet processing logic to the network, ( ii ) by using smart tagging to setup and exploit the affinity of traffic classes, and ( iii ) by using tag-based hardware dispatching to carry out the remaining packet processing at the speed of the servers’ cores, with zero inter-core communication. Moreover, Metron transparently integrates, manages, and load balances proprietary “blackboxes” together with Metron service chains. Metron realizes stateful network functions at the speed of 100GbE network cards on a single server, while elastically and rapidly adapting to changing workload volumes. Our experiments demonstrate that Metron service chains can coexist with heterogeneous blackboxes, while still leveraging Metron’s accurate dispatching and load balancing. In summary, Metron has ( i ) 2.75–8× better efficiency, up to ( ii ) 4.7× lower latency, and ( iii ) 7.8× higher throughput than OpenBox, a state-of-the-art NFV system.

show abstract

What You Need to Know About (Smart) Network Interface Cards

Cited by 14 publications

References 31 publications

A Survey on In-Network Computing: Programmable Data Plane and Technology Specific Applications

A Survey on In-Network Computing: Programmable Data Plane and Technology Specific Applications

High-speed Connection Tracking in Modern Servers

Metron

Contact Info

Product

Resources

About