BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20250309T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20241103T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20250215T232356Z UID:25E30DA2-8D26-412F-9D30-C04E20C99CFD DTSTART;TZID=America/New_York:20250212T183000 DTEND;TZID=America/New_York:20250212T200000 DESCRIPTION:Abstract: Almost all traffic on the Internet today is sent by C ongestion Control Algorithms\, which aim to maximize utilization of availa ble bandwidth on an Internet link while simultaneously sharing this bandwi dth equally with competing traffic. With the rise of heterogeneous congest ion control algorithms and increasingly complex application control loops (e.g. adaptive bitrate algorithms found in video streaming)\, the Internet community has expressed growing concern that network bandwidth allocation s are unfairly skewed\, and that some Internet services are ‘winners’ at the expense of ‘losing’ services when competing over shared bottlen eck links. In this paper\, we provide the first study of fairness between live\, end-to-end services with distinct workloads. Put simply\, if you an d your roommate are watching Netflix and YouTube on a bandwidth-constraine d Internet link\, would you end up streaming your video at the lowest reso lution while your roommate enjoys a high quality 4K stream\, or would the outcome be fairer? Among our findings\, we observe that services typically achieve less-than-fair outcomes: on average\, the ‘losing’ service ac hieves only 72% of its max-min fair share of link bandwidth. We also find that some services are significantly more contentious than others: for exa mple\, one popular file distribution service causes competing applications to obtain as low as 16% of their max-min fair share of bandwidth when com peting in a moderately-constrained setting.\n\nSpeaker(s): \, Adithya \n\n Room: Room 316\, Bldg: 3rd floor theater space \, 135 N. Bellefield\, Pitt sburgh\, Pennsylvania\, United States LOCATION:Room: Room 316\, Bldg: 3rd floor theater space \, 135 N. Bellefiel d\, Pittsburgh\, Pennsylvania\, United States ORGANIZER:bpalan@pitt.edu SEQUENCE:17 SUMMARY:Prudentia: Findings of an Internet Fairness Watchdog URL;VALUE=URI:https://events.vtools.ieee.org/m/463025 X-ALT-DESC:Description: <br /><p class="MsoNormal" style="text-align: justi fy\;"><strong><em><span style="mso-bidi-font-size: 12.0pt\; mso-fareast-fo nt-family: Aptos\; mso-font-kerning: 1.0pt\; mso-ligatures: standardcontex tual\;">Abstract: </span></em></strong><span style="mso-bidi-font-size: 12 .0pt\;">Almost all traffic on the Internet today is sent by Congestion Con trol Algorithms\, which aim to maximize utilization of available bandwidth on an Internet link while simultaneously sharing this bandwidth equally w ith competing traffic. With the rise of heterogeneous congestion control a lgorithms and increasingly complex application control loops (e.g. adaptiv e bitrate algorithms found in video streaming)\, the Internet community ha s expressed growing concern that network bandwidth allocations are unfairl y skewed\, and that some Internet services are &lsquo\;winners&rsquo\; at the expense of &lsquo\;losing&rsquo\; services when competing over shared bottleneck links. In this paper\, we provide the first study of fairness b etween live\, end-to-end services with distinct workloads. Put simply\, if you and your roommate are watching Netflix and YouTube on a bandwidth-con strained&nbsp\;Internet link\, would you end up streaming your video at th e lowest resolution while your roommate enjoys a high quality 4K stream\,& nbsp\;or would the outcome be fairer? Among our findings\, we observe that services typically achieve less-than-fair outcomes: on average\, the &lsq uo\;losing&rsquo\; service achieves only 72% of its max-min fair share of link bandwidth. We also find that some services are significantly more con tentious than others: for example\, one popular file distribution service causes competing applications to obtain as low as 16% of their max-min fai r share of bandwidth when competing in a moderately-constrained setting.</ span></p> END:VEVENT END:VCALENDAR