Better never than late: meeting deadlines in datacenter networks

By Christo Wilson, Hitesh Ballani, Thomas Karagiannis, Antony Ian Taylor Rowstron. Conference on Applications, Technologies, Architectures, and Protocols for Computer Communication, 2011

715 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Information Systems

View Paper: https://www.semanticscholar.org/paper/6c0c61afc3e65b6a23806d8eb6f3f9e1bfe15990

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Information Systems

Citations in the corpus

pFabric: minimal near-optimal datacenter transport (2013, background): “D3 [21] first proposed using deadline information in conjunction with explicit rate control to associate rates to flows.”
Efficient Coflow Scheduling Without Prior Knowledge (2015, neutral): “, PDQ [29], pFabric [10], and D(3) [47]) and non-clairvoyant (e.”
A Survey on Network Methodologies for Real-Time Analytics of Massive IoT Data and Open Research Issues (2017, background): “Readers may refer to the results in [167] that demonstrate the ratio of successful packets deliv-”
Data Center Network Virtualization: A Survey (2013, background): “For example, S 3 [64] is a ﬂow control mechanism that aims to meet ﬂow deadlines.”
Low Latency Geo-distributed Data Analytics (2015, background): “3) Optimizing communication patterns: Flow schedulers like D3 [55], PDQ [35], DeTail [62], and D2TCP [51] aim to improve flow completion times or guarantee deadlines.”
Mirror mirror on the ceiling: flexible wireless links for data centers (2012, background): “We note that more complex policies can be added to our scheduler, such as alternative ranking metrics that prioritize job by their deadlines in deadline-driven data centers [43].”
TIMELY: RTT-based Congestion Control for the Datacenter (2015, background): “Since traditional loss-based transports do not meet these strict requirements, new datacenter transports [10,18,30,35, 37,47], take advantage of network support to signal the on-set of congestion (e.g., DCTCP [35] and its successors use ECN), introduce ﬂow abstractions to minimize completion…”
Heracles: Improving resource efficiency at scale (2015, background): “There are also several networking protocols that prioritize short messages for LC tasks over large messages for BE tasks [5, 76].”
Achieving high utilization with software-driven WAN (2013, background): “An example is when one DC contacts another in the process of responding to a user request because not all information is available in the ﬁrst DC. Interactive traﬃc is highly sensitive to loss and delay; even small increases in response time (100 ms) degrade user experience [35].”
M2R: Enabling Stronger Privacy in MapReduce Computation (2015, background): “We assume that the network round trip latency to the client is only 1ms — an optimistic approximation since the average round trip delay in the same data center is 10 − 100ms [4, 61].”
Scalable Network Virtualization in Software-Defined Networks (2013, methodology): “Experiments with our FlowN prototype, built as an extension to the NOX OpenFlow controller, show that our solution scales to a large number of tenants.”
Finishing flows quickly with preemptive scheduling (2012, methodology): “D(3): We implemented a complete version of D(3) [20], including the rate request processing procedure, the rate adaptation algorithm (with the suggested parameters α = 0.”
Homa: a receiver-driven low-latency transport protocol using network priorities (2018, methodology): “Recent years have seen numerous proposals for better transport protocols, including improvements to TCP [2, 3, 31] and a variety of new protocols [4, 6, 14, 15, 17, 25, 32].”
Fastpass (2014, background): “Schemes such as pFabric [8] and PDQ [20] modify the switches to reduce ﬂow completion time, while D 3 [36] and PDQ minimize ﬂow completion times to meet deadlines.”
Participatory networking: an API for application control of SDNs (2013, background): “Deadlines can be communicated to supporting routers such as those proposed in D3 [48].”
DeTail: reducing the flow completion time tail in datacenter networks (2012, background): “4 Current Solutions Insufficient DCTCP and D [13, 33] are two single path solutions recently proposed to reduce the completion times of latencysensitive flows.”
Resource Management in Clouds: Survey and Research Challenges (2015, background): “[179] and D(3), a deadline-aware control protocol [241].”
Decentralized task-aware scheduling for data center networks (2014, methodology): “The ﬁrst two workloads are based on the Bing and Facebook traces discussed earlier ( § 2) while the third one models a more homogeneous application with ﬂow sizes that are uniformly distributed across [2KB, 50KB] (as suggested in prior work [4, 28, 17]).”