Skip to main content

Collected Datasets

Here you can find the datasets containing the latency measurements we performed in several Cloud Computing platforms. We obtained the data by running multiple instances of our measurement agent in Linux VMs placed in different datacenters within a vendor-specific platform and recording the results.

Amazon Web Services (AWS)

Between December 2016 and February 2017 we ran a campaign of measurements to analyze the communication latency between software components running in the main datacenters of the Amazon Web Services Cloud Computing platform.

More specifically, the dataset contains a set of latency measurements campaigns, which in turn were performed between the following datacenters. Here is the map between the filenames where data related to a single campaign was recorded and the corresponding AWS datacenters:

File                                Datacenters
casi-ping.2016.12.02-11.08.53.txt   US West (N. California) to Asia Pacific (Singapore) 
irsi-ping.2017.02.16-08.38.35.txt   EU (Ireland) to Asia Pacific (Singapore) 
orsi-ping.2016.12.02-11.14.44.txt   US West (Oregon) to Asia Pacific (Singapore) 
orsy-ping.2016.12.02-11.15.43.txt   US West (Oregon) to Asia Pacific (Sydney) 
orto-ping.2016.12.02-11.12.52.txt   US West (Oregon) to Asia Pacific (Tokyo) 
sasi-ping.2016.12.02-11.16.49.txt   South America (São Paulo) to Asia Pacific (Singapore) 
sisy-ping.2016.12.02-11.13.50.txt   Asia Pacific (Singapore) to Asia Pacific (Sydney) 
sito-ping.2016.12.02-11.18.12.txt   Asia Pacific (Singapore) to Asia Pacific (Tokyo) 
syvi-ping.2016.12.02-11.19.11.txt   Asia Pacific (Sydney) to US East (N. Virginia) 

You can download the entire dataset, containing information pertaining to all the measurmente campaigns reported above, here (please note that the dataset size is slightly less than 50 MiB).

IBM Cloud

Coming soon.

Google Cloud Platform

Coming soon. 

Microsoft Azure

Coming soon.

Processing the datasets

We created a set of Bourne shell scripts that (heavily leveraging on AWK and grep) extract specific information from one of the latency measurement campaigns between two datacenters contained in the datasets above.

More specifically: the getPingStats.sh script calculates aggregated statistics for a single campaign, the getRTT.sh script extracts only RTT information from a single campaign, the getTTL-RTT.sh script extracts both RTT and TTL information from a single campaign, and the getTTL-RTT-vs-seq.sh extracts RTT, TTL, and sequence number information from a single campaign.

Here you can download all the above mentioned scripts in a single ZIP file.






Comments