Tales of the Tail: Hardware, OS, and Application-level Sources of Tail Latency
The DOI matches its Crossref record; 24 of the 24 references it cites were found in a registry or an index.
Abstract
Interactive services often have large-scale parallel implementations. To deliver fast responses, the median and tail latencies of a service's components must be low. In this paper, we explore the hardware, OS, and application-level sources of poor tail latency in high throughput servers executing on multi-core machines.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
MapReduce: simplified data processing on large clusters
Dean, J., & Ghemawat, S. (2008) · Communications of the ACM · doi:10.1145/1327452.1327492 · Cited by 18,476 · cited 3×
“Related techniques have been used to avoid the “straggler” problem in data-parallel computation frameworks like MapReduce and Spark [2, 6, 9, 30, 31].”
MethodResilient distributed datasets: a fault-tolerant abstraction for in-memory cluster computing
Zaharia, M., Chowdhury, M., Das, T., Dave, A., Ma, J. H., McCauley, M., Franklin, M. J. M., Shenker, S. J., & Stoica, I. G. (2012) · UC Berkeley · Cited by 4,891
Lamport, L. (1998) · ACM Transactions on Computer Systems · doi:10.1145/279227.279229 · Cited by 2,737 · cited 1×
“” The techniques they used include re-issuing slow requests to a different host [6, 22], replicating data using quorum protocols that do not require every replica to answer [11], or accepting slightly incomplete results in information retrieval systems.”
MethodDean, J., & Barroso, L. A. (2013) · Communications of the ACM · doi:10.1145/2408776.2408794 · Cited by 1,852 · cited 3×
“Their goal is to allow parallel systems to tolerate latency variability in individual components, to “create a predictably responsive whole out of less-predictable parts [5].”
BackgroundInfluentialImproving MapReduce performance in heterogeneous environments
Zaharia, M., Konwinski, A., Joseph, A. D., Katz, R. H., & Stoica, I. G. (2008) · doi:10.5555/1855741.1855744 · Cited by 1,588
Nishtala, R., Fugal, H., Grimm, S., Kwiatkowski, M., Lee, H., Li, H. C., McElroy, R., Paleczny, M., Peek, D., Saab, P., et al. (2013) · Networked Systems Design and Implementation · Cited by 978
Reining in the outliers in map-reduce clusters using Mantri
Ananthanarayanan, G., Kandula, S., Greenberg, A. G., Stoica, I. G., Lu, Y., Saha, B., & Harris, E. F. (2010) · Operating Systems Design and Implementation · doi:10.5555/1924943.1924962 · Cited by 636 · cited 2×
“Related techniques have been used to avoid the “straggler” problem in data-parallel computation frameworks like MapReduce and Spark [2, 6, 9, 30, 31].”
MethodBetter never than late: meeting deadlines in datacenter networks
Wilson, C., Ballani, H., Karagiannis, T., & Rowtron, A. (2011) · Proceedings of the ACM SIGCOMM 2011 conference · doi:10.1145/2018436.2018443 · Cited by 579 · cited 2×
“The network can also be an important source of tail latency in datacenters, as has been studied extensively by previous work [1, 15, 19], but this is outside the scope of our study.”
Background~ " " " ' l I ~ " " -" . : -· " J
Cox, G., & Langford, B. (2006) · doi:10.4324/9780203973493-196 · Cited by 501
Deadline-aware datacenter tcp (D2TCP)
Vamanan, B., Hasan, J., & Vijaykumar, T. N. (2012) · Proceedings of the ACM SIGCOMM 2012 conference on Applications, technologies, architectures, and protocols for computer communication · doi:10.1145/2342356.2342388 · Cited by 473 · cited 2×
“The network can also be an important source of tail latency in datacenters, as has been studied extensively by previous work [1, 15, 19], but this is outside the scope of our study.”
Background
Cited by
Gan, Y., Zhang, Y., Cheng, D., Shetty, A., Rathi, P., Katarki, N., Bruno, A., Hu, J., Ritchken, B., Jackson, B., et al. (2019) · Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems · doi:10.1145/3297858.3304013 · Cited by 665
“We additionally track the time spent processing network requests, as opposed to application computation using a similar methodology to [58].”
MethodPARTIES: QoS-Aware Resource Partitioning for Multiple Interactive Services
Chen, S., Delimitrou, C., & Martínez, J. F. (2019) · Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems · doi:10.1145/3297858.3304005 · Cited by 267
ZygOS: Achieving Low Tail Latency for Microsecond-scale Networked Tasks
Prekas, G., Kogias, M., & Bugnion, E. (2017) · Proceedings of the 26th Symposium on Operating Systems Principles · doi:10.1145/3132747.3132780 · Cited by 207
Zhang, Y., Laurenzano, M. A., Mars, J., & Tang, L. (2014) · 2014 47th Annual IEEE/ACM International Symposium on Microarchitecture · doi:10.1109/micro.2014.53 · Cited by 163
Missing the Forest for the Trees: End-to-End AI Application Performance in Edge Data Centers
Richins, D., Doshi, D., Blackmore, M., Thulaseedharan Nair, A., Pathapati, N., Patel, A., Daguman, B., Dobrijalowski, D., Illikkal, R., Long, K., et al. (2020) · 2020 IEEE International Symposium on High Performance Computer Architecture (HPCA) · doi:10.1109/hpca47549.2020.00049 · Cited by 23
“Other work has broken down the end-to-end latency of requests, at various levels of granularity, ranging from evaluation of Internet speed and programming language patterns to operating system scheduling and memory latency [63, 64].”
Background
Metadata
- Authors
- Li, J., Sharma, N. K., Ports, D. R. K., & Gribble, S. D.
- Venue
- Proceedings of the ACM Symposium on Cloud Computing
- Publisher
- ACM
- Published
- 2014-11-03
- Pages
- 1-14
- OpenAlex
- W1994618472
- Cited by
- 235
- References
- 24
- FWCI
- 14.78
- Type
- conference paper
- Language
- en
- Open access
- closed
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.