BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260202T201806Z
LOCATION:266
DTSTART;TZID=America/Chicago:20251116T172000
DTEND;TZID=America/Chicago:20251116T173500
UID:submissions.supercomputing.org_SC25_sess221_ws_indis115@linklings.com
SUMMARY:From Edge to HPC: Investigating Cross-Facility Data Streaming Arch
 itectures
DESCRIPTION:Anjus George, Michael Brim, Christopher Zimmer, David Rogers, 
 Sarp Oral, and Zach Mayes (Oak Ridge National Laboratory (ORNL))\n\nIn thi
 s paper, we investigate three cross-facility data streaming architectures,
  Direct Streaming (DTS), Proxied Streaming (PRS), and Managed Service Stre
 aming (MSS). We examine their architectural variations in dataflow paths a
 nd deployment feasibility, and detail their implementation using the DS2HP
 C architectural framework and the SciStream memory-to-memory streaming too
 lkit on the production-grade ACE infrastructure at OLCF. We present a work
 flow-specific evaluation of these architectures using three synthetic work
 loads derived from the streaming characteristics of scientific workflows. 
 Through simulated experiments, we measure streaming throughput, round-trip
  time, and overhead under work sharing, work sharing with feedback, and br
 oadcast and gather messaging patterns commonly found in AI-HPC communicati
 on motifs. Our study shows that DTS offers a minimal-hop path, resulting i
 n higher throughput and lower latency, whereas MSS provides greater deploy
 ment feasibility and scalability across multiple users but incurs signific
 ant overhead. PRS lies in between, offering a scalable architecture whose 
 performance matches DTS in most cases.\n\nRecording: Livestreamed, Recorde
 d\n\nRegistration Category: Technical Program Reg Pass, Workshop Reg Pass\
 n\nSession Chairs: Ezra Kissel (Energy Sciences Network (ESnet)); Rafael S
 ilva-Guimaraes (Federal Institute of Espírito Santo, Department of Informa
 tics); Nik Sultana (Illinois Institute of Technology, SCinet); Cees de Laa
 t (University of Amsterdam, Lawrence Berkeley National Laboratory (LBNL));
  Sonja Filiposka (Ss. Cyril and Methodius Univeristy in Skopje (UKIM) Nort
 h Macedonia); and Akbar Kara (Ciena Corporation, SCinet)\n\n
END:VEVENT
END:VCALENDAR