PMNet: In-Network Data Persistence

Seemakhupt, Korakit; Liu, Sihang; Senevirathne, Yasas; Shahbaz, Muhammad; Khan, Samira

doi:10.1109/isca52012.2021.00068

Cited by 5 publications

(2 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data persistence is a necessity within IoT networks, as lost packets must have a way to be recovered, especially in systems where edge nodes are frequently unavailable due to being out of range or in a sleep state. An FPGA middleware server was designed for this purpose, which allows for persistent data within a network [ 80 ]. This middleware server acts as a mediator between a server and edge nodes, decoupling the server from the critical data path: when a node sends a request to the server, it is intercepted by the middleware, which then sends an ACK signal back to the node.…”

Section: Mid-range Iot Solutionsmentioning

confidence: 99%

Review of State-of-the-Art FPGA Applications in IoT Networks

Magyari

Chen

2022

Sensors

View full text Add to dashboard Cite

Modern networks used for integrating custom Internet of Things (IoT) systems and devices have restrictions and requirements unique to their individual applications. These application specific demands require custom designed hardware to maximize throughput, security and data integrity whilst minimizing latency, power consumption, and form factor. Within this paper, we describe current, state-of-the-art works that utilize FPGAs for IoT network developments. We analyze two categories of works: those that prioritize reducing power consumption, and those that prioritize networking features. Further, we describe how future works can improve upon these designs and therefore improve the efficiency of resource-constrained IoT networks.

show abstract

Section: Mid-range Iot Solutionsmentioning

confidence: 99%

Review of State-of-the-Art FPGA Applications in IoT Networks

Magyari

Chen

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…In modern datacenters, commodity networking devices, especially programmable switches and Smart NICs, have been leveraged to accelerate datacenter applications. Such approaches include caching [73,94,102,148], compute offloading [7,16,56,93,94,109,144,145,150], protocol offloading [34, 35, 72, 84, 91, 95-97, 138, 147, 161, 174, 181], load balancing [81,104], etc. However, due to the limited memory capacity of such devices (e.g., O(10MB) of on-chip SRAM for programmable switches [1,83] and O(10GB) of on-board DRAM for Smart NICs [101,122]), they may fall short of efficiently handling datacenter applications requiring large memory footprints.…”

Section: Related Workmentioning

confidence: 99%

ORCA: A Network and Architecture Co-design for Offloading us-scale Datacenter Applications

Yang¹,

Huang²,

Sun³

et al. 2022

Preprint

View full text Add to dashboard Cite

Responding to the "datacenter tax" and "killer microseconds" problems for datacenter applications, diverse solutions including Smart NIC-based ones have been proposed. Nonetheless, they often suffer from high overhead of communications over network and/or PCIe links. To tackle the limitations of the current solutions, this paper proposes ORCA, a holistic network and architecture co-design solution that leverages current RDMA and emerging cache-coherent off-chip interconnect technologies. Specifically, ORCA consists of four hardware and software components: (1) unified abstraction of inter-and intra-machine communications managed by one-sided RDMA write and cache-coherent memory write; (2) efficient notification of requests to accelerators assisted by cache coherence; (3) cache-coherent accelerator architecture directly processing requests received by NIC; and (4) adaptive device-tohost data transfer for modern server memory systems consisting of both DRAM and NVM exploiting state-of-the-art features in CPUs and PCIe. We prototype ORCA with a commercial system and evaluate three popular datacenter applications: in-memory key-value store, chain replication-based distributed transaction system, and deep learning recommendation model inference. The evaluation shows that ORCA provides 30.1∼69.1% lower latency, up to 2.5× higher throughput, and ∼ 3× higher power efficiency than the current state-of-the-art solutions.

show abstract

Holistic In-Network Acceleration for Heavy-Tailed Storage Workloads

Kim

2023

IEEE Access

View full text Add to dashboard Cite

PMNet: In-Network Data Persistence

Cited by 5 publications

References 61 publications

Review of State-of-the-Art FPGA Applications in IoT Networks

Review of State-of-the-Art FPGA Applications in IoT Networks

ORCA: A Network and Architecture Co-design for Offloading us-scale Datacenter Applications

Holistic In-Network Acceleration for Heavy-Tailed Storage Workloads

Contact Info

Product

Resources

About