Measuring and Evaluating TCP Splitting for Cloud Services

Pathak, Abhinav; Wang, Y. Angela; Huang, Cheng; Greenberg, Albert; Hu, Y. Charlie; Kern, Randy; Li, Jin; Ross, Keith W.

doi:10.1007/978-3-642-12334-4_5

Cited by 27 publications

(38 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although Google still directs the vast majority of client prefixes to servers in its own ASes, it has begun directing an additional 8% of them to servers off its network, representing a 393% increase in the number served from outside the network. By installing servers inside client ISPs, Google allows clients in these ISPs to terminate their TCP connections locally (likely at a satellite server that proxies requests to a datacenter [7,11,28], as it is extremely unlikely that Google has sufficient computation in these locations to provide its services). We perform reverse DNS lookups on the IP addresses of all front-ends we located outside of Google's network.…”

Section: Characterizing the Expansionmentioning

confidence: 99%

“…The fraction of client prefixes within 500km of their front-ends has increased from 21% to 89%, and the fraction within 1000km increased from 36% to 96%. Because many of the newer front-ends seem to be satellites that likely proxy traffic back to datacenters, it is hard to know the impact that decreasing the distance from client to front-end will have on application performance [28].…”

Section: Characterizing the Expansionmentioning

confidence: 99%

“…These consist of decentralized serving sites that contain one or more front-end servers. Clients of these infrastructures are directed to nearby front-ends, which either directly serve static content (e.g., video or images from a content distribution network like Akamai), or use split TCP connections to relay web acccess requests to back-end datacenters (e.g., Google's search infrastructure) [7,11,28].…”

Section: Introductionmentioning

confidence: 99%

“…An explicit non-goal of this work is to estimate the increase in Google's serving capacity: in placing front-ends in ISPs around the world, Google's expansion presumably focused on improving the latency of Web accesses through split-TCP connections [7,11,28], so proximity of front-ends to clients (this paper), and good path performance between clients and front-ends (future work) were more important than capacity increases.…”

Section: Introductionmentioning

confidence: 99%

“…To do so, providers direct clients to serving sites in or near the clients' networks. A client's TCP connection terminates at a front-end server in the serving site, but the front-end proxies the request back to one of the provider's large datacenters [28]. This arrangement has a number of potential advantages compared to directing the client directly to the datacenter.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Mapping the expansion of Google's serving infrastructure

Calder

Fan

et al. 2013

Proceedings of the 2013 Conference on Internet Measurement Conference

153

120

View full text Add to dashboard Cite

Modern content-distribution networks both provide bulk content and act as "serving infrastructure" for web services in order to reduce user-perceived latency. Serving infrastructures such as Google's are now critical to the online economy, making it imperative to understand their size, geographic distribution, and growth strategies. To this end, we develop techniques that enumerate IP addresses of servers in these infrastructures, find their geographic location, and identify the association between clients and clusters of servers. While general techniques for server enumeration and geolocation can exhibit large error, our techniques exploit the design and mechanisms of serving infrastructure to improve accuracy. We use the EDNS-client-subnet DNS extension to measure which clients a service maps to which of its serving sites. We devise a novel technique that uses this mapping to geolocate servers by combining noisy information about client locations with speed-of-light constraints. We demonstrate that this technique substantially improves geolocation accuracy relative to existing approaches. We also cluster server IP addresses into physical sites by measuring RTTs and adapting the cluster thresholds dynamically. Google's serving infrastructure has grown dramatically in the ten months, and we use our methods to chart its growth and understand its content serving strategy. We find that the number of Google serving sites has increased more than sevenfold, and most of the growth has occurred by placing servers in large and small ISPs across the world, not by expanding Google's backbone.

show abstract

Section: Characterizing the Expansionmentioning

confidence: 99%

Section: Characterizing the Expansionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Mapping the expansion of Google's serving infrastructure

Calder

Fan

et al. 2013

Proceedings of the 2013 Conference on Internet Measurement Conference

153

120

View full text Add to dashboard Cite

show abstract

Dissecting persistent instability of web service: A joint perspective of server schedule dynamics and path latency

Xue

Wang

Liu

et al. 2017

Int J Communication

View full text Add to dashboard Cite

Web service has become very popular because of its portal role like searching and video streaming, as well as its convenience of building. Complex communication and computation infrastructures behind modern web service are brought forward to satisfy diverse requirements, such as content delivery networks and clouds. However, no longitudinal studies about web serving pattern dynamics have been performed. To investigate the joint communication and computation patterns, we conducted a 1-year fine-grained measurement on world and China top 500 websites.Specifically, we examined path latency (perspective of communication) combined with hosting server dynamics (perspective of computation) from a longitudinal view of end user. Our result shows that persistent latency instability could exist with 2 contributors. One is persistent diurnal pattern, which exhibits significant daily latency discrepancy from 60 to 360 ms and can exist up to nearly 120 days out of 1 year. Another factor is dynamic server schedule. We design metrics and estimate to what extent server dynamics takes on over a long-term period and how it correlates with end-to-end latency. Further, we take a look at topological and geographical factors of server changes that affect path latency instability. The result shows that world and China top sites have different spatial dominators in terms of server change-related latency instability. KEYWORDS dynamic server scheduling, latency stability, longitude analysis, network measurement, web service Int J Commun Syst. 2018;31:e3441.wileyonlinelibrary.com/journal/dac persistent instability of web service: A joint perspective of server schedule dynamics and path latency. Int J Commun Syst. 2018;31:e3441. https://doi.

show abstract