Analysis of the Effect of Core Affinity on High-Throughput Flows

Hanford, Nathan; Ahuja, Vishal; Farrens, Matthew; Ghosal, Dipak; Balman, Mehmet; Pouyoul, Eric; Tierney, Brian

doi:10.1109/ndm.2014.10

Cited by 7 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To hide the latency of the main memory access, hierarchical cache subsystems have been proposed and utilized on multi-core systems [23], [24]. Network processors designed for high speed PPAs always rely extremely on cache technologies owing to the restricted CPU time budget per packet [25], [26]. A typical cache hierarchy may include:…”

Section: Stage Processing Time Model For Mplmentioning

confidence: 99%

Protocol-Aware Packet Scheduling Algorithm for Multi-Protocol Processing in Multi-Core MPL Architecture

Zhang

Wang

Sheng

et al. 2017

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYCache affinity has been proved to have great impact on the performance of packet processing applications on multi-core platforms. Flow-based packet scheduling can make the best of data cache affinity with flow associated data and context structures. However, little work on packet scheduling algorithms has been conducted when it comes to instruction cache (I-Cache) affinity in modified pipelining (MPL) architecture for multi-core systems. In this paper, we propose a protocol-aware packet scheduling (PAPS) algorithm aiming at maximizing I-Cache affinity at protocol dependent stages in MPL architecture for multi-protocol processing (MPP) scenario. The characteristics of applications in MPL are analyzed and a mapping model is introduced to illustrate the procedure of MPP. Besides, a stage processing time model for MPL is presented based on the analysis of multi-core cache hierarchy. PAPS is a kind of flowbased packet scheduling algorithm and it schedules flows in consideration of both application-level protocol of flows and load balancing. Experiments demonstrate that PAPS outperforms the Round Robin algorithm and the HRW-based (HRW) algorithm for MPP applications. In particular, PAPS can eliminate all I-Cache misses at protocol dependent stage and reduce the average CPU cycle consumption per packet by more than 10% in comparison with HRW.

show abstract

Section: Stage Processing Time Model For Mplmentioning

confidence: 99%

Protocol-Aware Packet Scheduling Algorithm for Multi-Protocol Processing in Multi-Core MPL Architecture

Zhang

Wang

Sheng

et al. 2017

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

“…al [5] describe their experience with long-haul 100Gbps networks, particularly talking about application design issues, protocol shortcomings and tuning recommendations to enable high-rate long-haul data transfers. The authors in [31][32] particularly discuss the effect of CPU affinity on network processing performance, and provide recommendations on how to assign interrupts and user processes to cores in order to achieve higher performance.…”

Section: Related Workmentioning

confidence: 99%

“…Bresnahan et al compared the performance and system resources utilized for GridFTP-TCP and GridFTP-UDT transfers [10]. The experiments, which were conducted on the TeraGrid [31] network between Argonne and Oak Ridge national laboratories, showed that while UDT can achieve higher throughput than TCP can on a long-haul network, the CPU utilization for TCP transfers was lower, in the range of 30-50%, compared with that of UDT transfers, which was around 80%.…”

Section: Related Workmentioning

confidence: 99%

Long-haul secure data transfer using hardware-assisted GridFTP

Rashti

Sabin

Kettimuthu

2016

Future Generation Computer Systems

View full text Add to dashboard Cite

Extreme-scale scientific collaborations require high-performance wide-area end-to-end data transports to enable fast and secure transfer of high data volumes among collaborating institutions. GridFTP is the de facto protocol for large-scale data transfer in science environments. Existing predominant network transport protocols such as TCP have serious limitations that consume significant CPU power and prevent GridFTP from achieving high throughput on long-haul networks with high latency and potential packet loss, reordering and jitter. On the other hand, protocols such as UDT that address some of the TCP shortcomings demand high computing resources on data transfer nodes. These limitations have caused underutilization of existing high-bandwidth links in scientific and collaborative grids. To address this situation, we have enhanced Globus GridFTP, the most widely used GridFTP implementation, by developing transport offload engines such as UDT and iWARP on SmartNIC, a programmable 10GbE network interface card (NIC). Our results show significant reduction in server utilization and full line-rate sustained bandwidth in high-latency networks, as measured for up to 100 ms of network latency. In our work, we also offload OpenSSL on SmartNIC to reduce host utilization for secure file transfers. The offload engine can provide line-rate data channel encryption/decryption on top of UDT offload without consuming additional host CPU resources. Lower CPU utilization leads to increased server capacity, which allows data transfer nodes to support higher network and data-processing rates. Alternatively, smaller or fewer DTNs can be used for a particular data rate requirement.

show abstract