Distributed Computing - What, why, how..
- 1. Distributed Computing What, why, and how Prepared by arinto <arinto (at) gmail (dot) com>
- 2. Outline ● What are DCs? ● DC in our daily life ○ Dropbox ○ Google Drive ○ Spotify ○ Parallel Processing ● DC in Indonesia ○ Energy ○ e-Government
- 3. What are DCs? Store and compute! Normal people usage: few hundreds of GB of data, browsing, email... Web companies and scientific communities : Petabytes of data!
- 4. Petabytes.. so what??!! a MS Word document : 40 KB = 40 kilobytes 1 megabyte = 2^10 kilobytes = 1024 kilobytes 1 gigabyte = 2^10 megabytes = 1024*1024 kilobytes > 1000000 kilobytes 1 terabyte = 2^10 gigabytes = 1024 gigabytes 1 petabyte = 2^10 terabytes = 1024*1024 gigabytes > 1000000 gigabytes > 1000000000000 kilobytes Google processed 24 Petabytes of data everyday in 2009!
- 5. Many as one
- 8. Scalability
- 9. Scalability
- 10. Sum them up! ● Many as one a.k.a transparent ○ User view them as a whole rather than a collection of cooperating components ● Resource sharing ○ Same set of resource shared by multiple users ● Fault tolerance ○ Still working despite failing ● Scalability ○ System or user can increase easily
- 11. Dropbox Store and retrieve data How do we transfer data ? Thumbdrive Backup? External HDD Stolen? Crashed?
- 12. Dropbox X ● Simple access from anywhere ● Simple sharing ● Back up
- 13. Dropbox X ● Simple access from anywhere ● Simple sharing ● Back up
- 14. Dropbox X DataCenter
- 15. Google Drive Write, store and read documents Collaboration ? Send back and forth.. MyDocuv223.doc Share? Send back and forth Free ? No!!!!! Can't access document from other device easily
- 16. Google Drive Access your data from anywhere anytime
- 17. Spotify Download, store and listen music How do we get the music ? Download! Legal? No!!! Clean? No!!!!! Quality? Baddd :( Can't access playlist from other device easily
- 18. Spotify Vision: Music service easier and better than illegal music!
- 19. Spotify Vision: Music service easier and better than illegal music!
- 20. Spotify Peer-to-peer with master backup! Vision: Music service easier and better than illegal music!
- 21. Parallel Processing a MS Word document : 40 KB = 40 kilobytes Google processed 24 Petabytes of data everyday in 2009! Estimated total document = 24*1000000000000 kilobytes / 40 kilobytes = 600000000000 If 1 docu needs 0.001 second, we need = 600000000 seconds = 10000000 hours = 416667 days = 1141.5 years!!!! Solution: parallelize!!
- 22. DC in Indonesia 1. Smart Grid a. Communication between Power Source b. Monitoring and providing appropriate response 2. Smart City and Internet of Things a. Building Management b. Interoperability 3. e-Government a. Open Data b. Online government service
- 23. Smart Grid Computing infrastructure, which is distributed! Ref: http://www.smartgrid.epri.com/Demo.aspx
- 24. Internet of Things (IoT) Nest Smart-Termostat ● Learn our behavior, adjust temperature accordingly ● Connected termostat to internet ● Web-based management portal Promotional video: http://www.nest.com/saving-energy/#videotour
- 25. IoT towards Smart City
- 26. IoT towards Smart City
- 27. IoT towards Smart City
- 28. e-Government - Open Data ● Open Data Stockholm (http://open.stockholm.se/oppna-data) - Population - Geodata, i.e: urban area, road network - Environmental Data - Traffic and Parking - School data ● Open Data Singapore (http://data.gov.sg) - Traffic and Parking
- 29. Conclusion 1. We know what Distributed Computing is! 2. Distributed Computing is real and used in daily life! 3. Some applications of Distributed Computing in our beloved country!