geocost: Cost-latency Optimized Multi-zone Cloud Storage Solution; Agile or not? Factors that Motivate Agile Development Method to Emerge and to be Popular

Liu, Haotian, School of Engineering and Applied Science, University of Virginia
Ku, Tsai-Hsuan, EN-Engineering and Society, University of Virginia
Shen, Haiying, EN-Comp Science Dept, University of Virginia

The theme of my STS research paper investigates the process Agile software development
method emerged and became popular in the past few decades with multiple data collection
approaches including in-person interviews with current software engineers of a diverse
background and past literature review, usage of some sociotechnical analysis frameworks and
my personal experience of leading a team of software engineers. In my technical project, I
participated in University of Virginia’s Pervasive Communication Lab and built geocost, a
reinforcement learning-based cost-latency optimized solution for cloud data storage.
Technical Report Summary
The main deliverable of my technical research project is geocost, an API-oriented module that
finds the balance between cost and data transaction latency while storing and serving data from
cloud storage provider such as AWS S3. The module consists of a forecast component which
utilized a RL model trained with historical cloud storage benchmarking data to make the most
cost-latency optimized decision and an server component based on fastAPI, a Python-based
ASGI server that works in sync with the forecast component to expose two API endpoints for
client to upload / download data to / from a certain cloud service provider.
STS Research Paper Summary
My STS research paper investigates the reason behind the emergence and success of Agile
software development method in the past few decades. By studying the industry-wide shift from
SaaP (Software as a Product) to SaaS (Software as a Service), the translation process in a
simulated software engineering network and interviewing with real software engineers of a
diverse background, the paper draws the conclusion that the emergence and popularity of Agile
development methodology can be attributed to shift of focus in consumer software requirement
and software development infrastructure.
Although seemingly not relevant at all, my STS research paper and technical report both shed
lights on the rapid evolution of today’s internet large-scale infrastructure. While my technical
reports explores a way of delivering customers the most cost-latency optimized solution of
storing and serving data from cloud storage service providers, my STS research paper uses these
cloud storage service providers as a prime example of SaaS and their emergence in the software
industry a driving force of Agile method’s adoption. Overall, the STS and technical research
both helped me gain insights of the software industry as a whole.

BS (Bachelor of Science)
Agile method, Cloud storage, geocost, AWS

School of Engineering and Applied Science
Bachelor of Science in Compute Science
Technical Advisor: Haiying Shen
STS Advisor: Tsai-Hsuan, Ku
Technical Team Members: Haotian Liu

Issued Date: