BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T143139Z
LOCATION:B312
DTSTART;TZID=America/New_York:20241118T153000
DTEND;TZID=America/New_York:20241118T160000
UID:submissions.supercomputing.org_SC24_sess811_ws_ciss105@linklings.com
SUMMARY:Protocol Buffer Deserialization DPU Offloading in the RPC Datapath
DESCRIPTION:Raphaël Frantz (Eindhoven University of Technology, Netherland
 s); Jerónimo Sánchez García (Aalborg University, Copenhagen); Marcin Copik
  (ETH Zürich); Idelfonso Tafur Monroy (Eindhoven University of Technology,
  Netherlands); and Juan José Vegas Olmos, Gil Bloch, and Salvatore Di Giro
 lamo (NVIDIA Corporation)\n\nIn the microservice paradigm, monolithic appl
 ications are decomposed into finer-grained modules invoked independently i
 n a data-flow fashion. The different modules communicate through remote pr
 ocedure calls (RPCs), which constitute a critical component of the infrast
 ructure. To ensure portable passage of RPC metadata, arguments, and return
  values between different microservices, RPCs involve serialization/deseri
 alization activities, part of the RPC data center tax. We demonstrate how 
 RPC server logic, including serialization/deserialization, can be offloade
 d to Data Processing Units (DPUs). This effectively reduces the RPC data c
 enter tax on the host, where applications' business logic runs. While we f
 ocus on offloading Protocol Buffers deserialization used by the popular gR
 PC framework, our findings can be applied to other RPC infrastructures. Ou
 r experimental results demonstrate that RPC offloading performs similarly 
 to traditional methods while significantly reducing CPU usage.\n\nTag: I/O
 , Storage, Archive\n\nRegistration Category: Workshop Reg Pass\n\nSession 
 Chairs: Glenn Brook (Cornelis Networks, University of Tennessee); Clayton 
 Hughes (Sandia National Laboratories); Nalini Kumar (Intel Corporation); H
 atem Ltaief (King Abdullah University of Science and Technology (KAUST)); 
 David Martin (Lawrence Berkeley National Laboratory (LBNL), Energy Science
 s Network (ESnet)); and Amit Ruhela (Texas Advanced Computing Center (TACC
 ), University of Texas)\n\n
END:VEVENT
END:VCALENDAR
