HNGraph: Parallel Graph Processing in Hybrid Memory Based NUMA Systems

Liu, Wei; Liu, Haikun; Liao, Xiaofei; Jin, Hai

doi:10.1109/cluster48925.2021.00063

Cited by 2 publications

(1 citation statement)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Future work in this area could address how edge nodes operating in Symmetric Multiprocessing [26], Massively Parallel Processing [27], Clustered Memory Scheduling [28] or Non-Uniform Memory Access [29], [30] systems could be synergised with cloud resources for increased efficiency in Big Data analytics.…”

Section: Discussionmentioning

confidence: 99%

Edge-Cloud Synergy: Unleashing the Potential of Parallel Processing for Big Data Analytics

Singh

Kiss

2022

2022 IEEE 13th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON)

View full text Add to dashboard Cite

If an edge-node orchestrator can partition Big Data tasks of variable computational complexity between the edge and cloud resources, major reductions in total task completion times can be achieved even at low Wide Area Network (WAN) speeds. The percentage time savings are greater with increasing task computational complexity and higher WAN speeds are required for low-complexity tasks. We demonstrate from numerical simulations that low-complexity tasks can benefit either by task partitioning between an edge node and multiple cloud servers. The orchestrator can also achieve greater time benefits by rerouting Big Data tasks directly to a single cloud resource if the balance of parameters (WAN speed and the ratio between edge and cloud processing speeds) is favourable.

show abstract