Big Brother Big Sister Bluemix Architecture from #HackathonCLT
•Download as ODP, PDF•
1 like•81 views
Big Brother and Big Sisters brought their enterprise system challenge to #HackathonCLT. I wanted to design a system that could support 10x the membership with the same expense. By using Bluemix to provide scale, the Watson APIs for both ingestion and analytics, and their Hyperledger implementation for security paired with HBase, I believe we have a potential solution. More to come!
Recommended
More Related Content
Similar to Big Brother Big Sister Bluemix Architecture from #HackathonCLT
Similar to Big Brother Big Sister Bluemix Architecture from #HackathonCLT (20)
More from Dave Callaghan
Recently uploaded
Recently uploaded (20)
Big Brother Big Sister Bluemix Architecture from #HackathonCLT
- 1. Big Brother Big Sisters Big Ideas Big Brother and Big Sisters needs a scalable, highly secure enterprise-grade system to ingest and analyze data. The goal is support 10x the children with the same expense ratio. Watson – Deep Analytics Blockchain – High security Bluemix – Highly scalable cloud infrasructure
- 2. So you know … Children and teens are often better at expressing themselves orally or verbally, not through a standardized form, Name Alberto Age 14 Gender M School James Woods Regional Hobbies soccer videogames
- 3. So you know … Adults are no better; possibly worse., Name Marcos Age 44 Gender M License Y Military Army Hobbies golf
- 4. How can we make more, better matches faster? The heart between the Big and the Little is Trust. There is already a system in place to do deep background checks, but trust is built on empathy. We need to learn about people as they are; not how they want others to see them. We need this system to be affordable, intuitive, capable of doing proportionally more with less volunteers and at the cutting
- 5. IBM Watson is a supercomputer combining Artificial Intelligence and machine learning for advanced human- computer interactions. Tradeoff Analytics Personality InsightsConversation Document Conversion
- 6. Tradeoff Analytics Personality Insights Conversation Derives insights from social media and transactional data to identify psychological traits for predictive behavior analytics. Help make complex decisions while taking into account multiple, often conflicting, goals Add chatbots for a natural language interface. Document Conversion Converts various types of documents into formats that can be understood by Watson services. Integration Analytics
- 7. Not only istheinsight much deeper, but it iscontinually updated based on social media, activity reportsand other resources. Thiscan identify and diagnoseissuesbeforethey progressfor deep safety.
- 8. We can move from a manual form-based matching system to continuous automated cognitive analytics. But how can security and privacy scale with such deep learning?
- 9. Blockchains use an open, distributed ledger that can record transactions between two parties efficiently and in a verifiable and permanent way through cryptography. There are many implementations and zero standards around blockchain. However, IBM’s Apache Hyperledger is the most enterprise friendly. The blockchain is a bad database, Blockchains are slow. Blockchains are cumbersome. Blockchains are resource intensive. But an ideal store of timestamped facts,
- 10. Hbase is a sparse, distributed, persistent multidimensional sorted map. Cells store data as key value pairs and provide tags that can be used for visibility based on the blockchain’s status. I kind of hacked the tag implementation, but ... Off-blockchain storage addresses many of the limitations of blockchain as a database.
- 11. So now we have a phenomenally secured system supporting an advanced AI platform. How can a non-profit with limited resources build, support and maintain a system like that?
- 12. Tip Gazing lovingly at your onsite servers is overrated. People who pay more for the cloud than onsite don’t know how to model cost properly. Doing things is terrible. Everything always breaks ... is awful. If it isn’t a differentiator, don’t spend scarce resources. Develop in containers, Implement with APIs, Deploy in pods, Run in the cloud.
- 13. So now we have a phenomenally secured system supporting an advanced AI platform that can be built, supported and maintained by low cost, open-source, Enterprise grade tools. How about a nerd-core name?
- 14. CHEWBACCA Stack Chatbots + Hbase + Watson + Blockchain + Containers + Cloud
- 15. CHEWBACCA Stack Chatbots + Hbase + Watson + Blockchain + Containers + Cloud Obviously
- 16. It’s more of a protocol droid’s stack, but still ... A combination of structured and natural language data is gathered by the chatbot and sent to the blockchain. Hyperledger stores the data in Hbase, keeping a hash pointer to the storage. Conversation Personality Insights Tradeoff Analytics