BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:MS Thesis Final Exam - Ashish Tulso Ramrakhiani
X-WR-TIMEZONE:Pacific Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260714T134338Z
UID:tag:localist.com\,2008:EventInstance_52551793159710
DTSTART:20260413T190000Z
DTEND:20260413T200000Z
DESCRIPTION:TITLE: EphFlow: Addressing Resource Limitations of Cross-platfo
 rm FaaS Workflows via Serverless Ephemeral VM Provisioning\n\nABSTRACT: Sc
 ientific workflows are essential for automating complex computational pipe
 lines. While serverless Function-as-a-Service (FaaS) cloud platforms have 
 the potential to enable wider adoption of workflows\, existing systems pre
 sent significant adoption barriers: vendor lock-in\, strict resource limit
 ations\, and limited language support. We present EphFlow\, a novel middle
 ware that enables serverless cross-platform execution of workflows across 
 multiple cloud platforms (without code modifications)\, exposing the FaaS 
 programming abstraction while supporting different languages\, and support
 ing dynamic instantiation of ephemeral VMs to host executions that exceed 
 the capacity of FaaS platforms. EphFlow introduces three key innovations: 
 (1) a client-server architecture that supports deployment of FaaS actions 
 in general-purpose containers\, abstracting serverless platform-specific A
 PIs while generalizing to multiple programming languages (Python\, R\, and
  Julia demonstrated)\; (2) automatic in-DAG provisioning of ephemeral VMs 
 that transparently overcomes FaaS execution time/memory limits by injectin
 g lifecycle management actions into workflow DAGs\; and (3) passive S3-bas
 ed coordination that eliminates the need for dedicated workflow engines. E
 xperiments demonstrate the ability to deploy workflows within or across fi
 ve different platforms\, building upon a common foundation of REST APIs fo
 r container invocation: GitHub Actions\, AWS Lambda\, Google Cloud Run\, O
 penWhisk\, and SLURM. The system has been qualitatively and quantitatively
  evaluated using both synthetic workflows and a realistic event-driven eco
 logical forecasting application (FLARE). Qualitative evaluations demonstra
 te the ability to deploy multi-language FaaS workflows across the above pl
 atforms and ephemeral in-workflow VM provisioning as GitHub self-hosted Ru
 nners on AWS EC2 without any code changes or managed servers. Event-driven
  deployment of a synthetic workflow across Lambda\, Google Cloud and GitHu
 b Actions over two weeks (1\,051 invocations) quantifies the total latency
  between container action invocations in a workflow DAG (with median rangi
 ng from 9.3s in Lambda to 131.0s with ephemeral VM provisioning) as well a
 s overheads associated with individual steps in the middleware’s entry p
 oint (0.3–5 seconds to process a workflow configuration payload and invo
 ke the user function). Experiments also demonstrate the ability to combine
  FaaS and ephemeral VM provisioning in a single workflow with actions that
  exceed AWS Lambda resource limits. These results show the feasibility of 
 deployment of FaaS workflows across serverless providers without vendor lo
 ck-in\, in particular for those with long-running actions (minutes or long
 er).\n\n \n\nMAJOR ADVISOR: Renato Figueiredo\nCOMMITTEE: Kyle Hale\nCOMMI
 TTEE: Wenqian Dong\nGCR: Adam Branscum
GEO:44.567164;-123.278692
LOCATION:Kelley Engineering Center\, 1007
SUMMARY:MS Thesis Final Exam - Ashish Tulso Ramrakhiani
URL;VALUE=URI:https://events.oregonstate.edu/event/ms-thesis-final-exam-ash
 ish-tulso-ramrakhiani
CATEGORIES:Lecture or Presentation
END:VEVENT
END:VCALENDAR