In modern science and engineering, it is essential to use data analysis and calculation using computers. However, large-scale computing is labor-intensive to achieve extensibility and manageability of computing resources. We are designing and implementing a cloud platform that standardizes development, running, and sharing of the data processing tasks using cloud technologies and Python 3. We are also adding research/education services on top of it.
In this talk, we are going to share what we have learned during 2 months of development experiences. In particular, it will include the architecture of our platform, experiences in the design and implementation process, and common caveats to care when you do a similar work. We hope to share our motivation that allowed our pathway over such a mine field with you.
Summary
Modern science and Python
Online programming playground for researchers and educators
Technical challenges
Sandboxing!
Resource consolidation
Fast uploads/downloads of data
Choice of tech stack
Advantages and disadvantages of Python
It's time to go Python 3
On-premise vs. Hosting vs. Cloud (AWS / Azure / GCE)
Docker containers 와 kubernetes
"Shoveling"
Warring states of Javascript frameworks + UI frontends
Entangling Waltz of Polymer, webcomponents, and Django
Backends: Mad-max around Docker
Security holes in ipython/Jupyter in the perspective of developer and hacker