A Switch Architecture Guaranteeing QoS Provision and HOL Blocking Elimination

Martínez, A.; García, Pedro J.; Alfaro, Francisco J.; Sánchez, José L.; Flich, José; Quiles, Francisco J.; Duato, J.

doi:10.1109/tpds.2008.62

Cited by 5 publications

(7 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…IBA defines a high-performance interconnection technology which provides low-latency and high-bandwidth, with low processing overhead. Among other features, IBA provides flexibility to implement any network topology, as well as mechanisms to configure efficient routing algorithms, guarantee deadlock-freedom [2], provide Quality-of-Service (QoS) [3] or congestion management [4]. Indeed, IBA is the dominant interconnect technology in the systems included in the Top500 list [5], reaching the 47.4% of the system included in this ranking in November 2015.…”

Section: Motivationmentioning

confidence: 99%

Combining OpenFabrics Software and Simulation Tools for Modeling InfiniBand-Based Interconnection Networks

Maglione-Mathey

Yébenes

Escudero-Sahuquillo

et al. 2016

2016 2nd IEEE International Workshop on High-Performance Interconnection Networks in the Exascale and Big-Data Era (HiPINEB)

View full text Add to dashboard Cite

The design of interconnection networks is becoming extremely important for High-Performance Computing (HPC) systems in the Exascale Era. Design decisions like the selection of the network topology, routing algorithm, fault tolerance and/or congestion control are crucial for the network performance. Besides, the interconnection network designers are also focused on creating middleware layers compatible to different network technologies, which make it possible for these technologies to interoperate. One example is the OpenFabrics Software (OFS) used in HPC for breakthrough applications that require high efficiency computing, wire-speed messaging, microsecond latencies and fast I/O for storage and file systems. OFS is compatible with several HPC interconnect technologies, like InfiniBand, iWarp or RoCE. One challenge in the design of new features for improving the interconnection network performance is to model in specific simulation tools the latency introduced by the OFS modules into the network traffic. In this paper, we present a work-in-progress methodology to combine the OFS middleware with OMNeT++based simulation tools, so that we can use some of the OFS modules, like OpenSM or ibsim, combined with simulation tools. We also propose a set of tools for analyzing the properties of different network topologies. Future work will consist on modeling other OFS modules functionality in network simulators.

show abstract

Section: Motivationmentioning

confidence: 99%

Combining OpenFabrics Software and Simulation Tools for Modeling InfiniBand-Based Interconnection Networks

Maglione-Mathey

Yébenes

Escudero-Sahuquillo

et al. 2016

2016 2nd IEEE International Workshop on High-Performance Interconnection Networks in the Exascale and Big-Data Era (HiPINEB)

View full text Add to dashboard Cite

show abstract

“…Taking all this into account, we believe that it is possible an integrated management of the available network resources, oriented to provide both QoS and Congestion Management by means of a clever use of the same set of resources. To the best of our knowledge, the only proposals that follow this integrated approach are those presented in [6] and [7]. However, both proposals require resources that are not available in current interconnect technologies, such as Content-Addressable Memories (CAMs) at each port, and other logic.…”

Section: Motivationmentioning

confidence: 99%

“…On the other hand, VLs [6][7] have only one entry in the table with a weight of 1, thus each one has been assigned a 5% of theoretical percentage of network bandwidth. Therefore, the total percentage of network bandwidth assigned to each SLP is: 60% for SLP0 (VLs[0-3]), 30% for SLP1 (VLs [4][5]), and 10% for SLP2 (VLs [5][6]).…”

Section: Experiments Definitionmentioning

confidence: 99%

“…Therefore, the total percentage of network bandwidth assigned to each SLP is: 60% for SLP0 (VLs[0-3]), 30% for SLP1 (VLs [4][5]), and 10% for SLP2 (VLs [5][6]). Note that all the entries of a given VL in the VLArbitrationTable have been scattered across the table in order to efficiently consume the bandwidth assigned to that VL (see [16] for more details).…”

Section: Experiments Definitionmentioning

confidence: 99%

“…This mechanism combines both QoS and BBQ for HoL-blocking avoidance, as explained in Section III. The queuing scheme BBQ is used for each one of the three SLPs: BBQ with 4VLs is applied to SLP0 (VLs[0-3]), and BBQ with 2VLs is used both for SLP1 (VLs [4][5]) and for SLP2 (VLs [6][7]). Thus the network is divided into 4 bands for SLP0 and into 2 bands for SLP1 and SLP2.…”

Section: Experiments Definitionmentioning

confidence: 99%

See 2 more Smart Citations

Combining HoL-blocking avoidance and differentiated services in high-speed interconnects

Yébenes

Escudero-Sahuquillo

Gómez

et al. 2014

2014 21st International Conference on High Performance Computing (HiPC)

View full text Add to dashboard Cite

Current high-performance platforms such as Datacenters or High-Performance Computing systems rely on highspeed interconnection networks able to cope with the everincreasing communication requirements of modern applications. In particular, in high-performance systems that must offer differentiated services to applications which involve traffic prioritization, it is almost mandatory that the interconnection network provides some type of Quality-of-Service (QoS) and CongestionManagement mechanism in order to achieve the required network performance. Most current QoS and Congestion-Management mechanisms for high-speed interconnects are based on using the same kind of resources, but with different criteria, resulting in disjoint types of mechanisms. By contrast, we propose in this paper a novel, straightforward solution that leverages the resources already available in InfiniBand components (basically Service Levels and Virtual Lanes) to provide both QoS and Congestion Management at the same time. This proposal is called CHADS (Combined HoL-blocking Avoidance and Differentiated Services), and it could be applied to any network topology. From the results shown in this paper for networks configured with the novel, cost-efficient KNS hybrid topology, we can conclude that CHADS is more efficient than other schemes in reducing the interferences among packet flows that have the same or different priorities.

show abstract

Buffering and prioritization in switches for fast processing table‐miss packets in software‐defined networks

Lai

Hailemariam

Chen

et al. 2021

Int J Communication

View full text Add to dashboard Cite

Summary In software‐defined networks, the switch forwards incoming packets according to forwarding rules recorded in the flow entries. When a switch receives a table‐miss packet, meaning no match with any flow entry, it sends this packet as a Packet‐In message to the controller for its further processing. Many Packet‐In messages will cause large overhead and long packet delay. This paper proposes a novel method, Packet‐In Buffering and Prioritization (PIBP), which buffers Packet‐In messages and prioritizes these messages to reduce the number of Packet‐In messages and accelerate their processing, respectively. The concept of PIBP is sending only the first table‐miss packet of each flow to the controller. The other table‐miss packets belonging to the same flow are temporarily stored in the switch. Moreover, these messages have a higher priority. That is, after the packet sent from the controller has been forwarded from the switch, the buffered packets belonging to the same flow are also immediately forwarded. We formally analyze the performance of PIBP with queuing theory. The analytical and simulation results show that PIBP can decrease the average delay of table‐miss packets compared with two typical methods, Priority Queue and Mismatched Packets Table.

show abstract

A Switch Architecture Guaranteeing QoS Provision and HOL Blocking Elimination

Cited by 5 publications

References 11 publications

Combining OpenFabrics Software and Simulation Tools for Modeling InfiniBand-Based Interconnection Networks

Combining OpenFabrics Software and Simulation Tools for Modeling InfiniBand-Based Interconnection Networks

Combining HoL-blocking avoidance and differentiated services in high-speed interconnects

Buffering and prioritization in switches for fast processing table‐miss packets in software‐defined networks

Contact Info

Product

Resources

About