Presenting the "Un-Apps"

Computer Science Alumni Lecture by Vance Pan at UBC on September 19, 2013 on the topic of the "Un-Apps". Abstract of the lecture:

"We believe that the Internet of Things will be enabled by a massive global network of small, agile, specific purpose programs that can capture the events and data that the real world generates. These programs will need to able to share their data with any other programs in the network. Programmers should be able to focus strictly on building the single purpose programs without constraints or dependencies of other programs or functionality (e.g. APIs). As no one program can understand everything, there will need to be a way of expressing how these different programs interact, like a road map that charts the course for the data as it moves from program to program. Data should not be "sucked into” applications and trapped there, they should be free.

This is what we call an Un-App. As typical applications perform a set of computations to provide the value of the application, an Un-App does the same thing with a road map. What is different is that with an Un-App, a program can belong to more than one roadmap. So while you would never think of being able to use live running logic from one traditional application in another application, with an Un-App it is a natural capability. With Un-Apps, any program you create can naturally communicate and integrate with any other program. Now imagine being able to aggregate small road maps into larger, more sophisticated road maps."

Transcript

1. Presenting the “Un-Apps” Vance Pan UBC CPSC Alumnus September 2013
2. Speaker Intro • Co-founder, Nodally Technologies Inc. • UBC CPSC, ICL MBA • Simba, Pivotal, BusinessObjects, SAP, DataCVG • Databases, CRM, BI, SaaS, PaaS • Marketing, Dev, QA, Doc, Support, PgM, PdM, Consulting, Pre-Sales, Alliance, Finance, Training • CPSC Mentor since 2010 • SAP Mentor (#58) since 2012 • 3+ Years in China
3. Agenda • Why is Software Flawed? • What is an Un-App? • Case Study • What Un-Apps look like • The Future of Un-Apps
4. Software is Flawed • What is an “application”? • A single program made up of many features • Why is today’s software “flawed”?
5. Root Problem with Applications • They’re monolithic • Hard to integrate, customize, sustain, re-use • Culprit: glue code • Turn into legacy systems / info silos • OOP, CBD, patterns, APIs help, but flaws remain
6. Introducing the “Un-Apps” Enabler of event-driven software systems by micronizing software and liberating data
7. The Un-Apps Design • Un-Apps are not monolithic, but aggregate of micro-apps running together • Logic and Data not trapped inside an app, glued together nor limited to one locality • Event-driven architecture (key to automating distributed software) • Bottom-up approach to creating solutions • Promote continuous improvements
8. “Micronizing software” micro-app: small, independent program written to perform a specific task where integration + rapid change + re-use part of its DNA
9. “Liberating Data” Platform enables orchestration and allows data to flow freely & securely anywhere
10. “Context by Proximity” data context available when the data arrives at a specific Un-App same data can flow to other Un-Apps where context can be completely different
11. Un-App = micro-app with logic + directions of flow of data + context
12. any running code can share data with any other live, running code anytime, anywhere ... and any app can share data with any other apps
13. Unix Pipe Analogy output of one process (ps) is sent to another process (grep) as input through the pipe command
14. same code can be used in parallel within any number of apps
15. App 1 App 2 App 4 App 3 X2C FW C2H FR GR DB RSS SMS GS
16. micronizing software + freeing data and logic + event-driven architecture enable an entirely new level of capabilities: reuse share extend customize maintain integrate mashup
17. “Un-App” = Bottom-up Approach Focus on solving specific problems without constraints of others Then augment, aggregate, and grow organically to form a perfect-fit solution
18. • Manufacturer of power solutions with overseas suppliers sending order updates over email • Takes 16 hrs/wk manually reviewing the updates, make calculations, and updating record in ERP, for 1 supplier • Time consuming and prone to delays + human errors • Typical solutions (e.g. EDI) too expensive and disruptive to existing processes • Un-Apps solution is a series of micro-apps for one supplier Case Study: Manufacturer A (1) EN EW EU OP SP < 1000 lines of code = 16 hrs/week productivity gain
19. • To create the same solution to work with a different supplier: • Reuse all micro-apps • Create new instance of Email Watcher and update configurations • Update spreadsheet parser to work with different schema Case Study: Manufacturer A (2) Hours of work = 16 hrs/week productivity gain
20. Un-Apps In Action
21. Internet of Things • What is the IoT? • “A world where all objects around us are connected wirelessly to the internet, enabled by a massive network of software apps” • “Internet of Everything”
23. To fully realize the vision of IoT, software needs to be event-driven and bottom-up Not possible with current paradigm and top-down approaches Role of Software in IoT
24. Un-Apps Solve Bigger Issues? • Problem: traffic accidents at an intersection • Devices: speed cameras, traffic lights, GPS • Challenges: devices can’t communicate; different domain expertise required for different devices
25. Community-Driven Solution Design • Un-Apps are Open Sourced • Language, platform, cloud agnostic platform • Social Coding: share, learn, teach, collaborate, crowdsource, feedback, discuss, have fun • Aggregate small solutions into bigger ones • Power of Community
26. Our Vision
27. About Nodally • Doug Squarebriggs (DougS@nodally.com) • Vance Pan (VanceP@nodally.com) • James Hsueh (JamesH@nodally.com) • Company site: www.nodally.com • Product site: www.xenograte.com • Twitter: https://twitter.com/nodally • Blog: http://blog.xenograte.com/ • Open Source Project: https://github.com/Nodally/xenograte-xct
28. Thank You