Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Cloud Computing ...changes everything
1.
2. Cloud Computing.... Virtualization Grid computing Application hosting Utility computing Platform as a service Infrastructure as a service Software as a service
3.
4. Analogy with the evolution of electrical power From each company generating their own power to a utility
14. Data Center Level Computing Google’s new data center on the Columbia river, Oregon Thousands and thousands of commodity parts built into a system to essentially serve a single application Power and Cooling major drivers of cost
15. Data + compute -> new opportunities 'Semi-structured' data emerges Mogile, Bigtable, Hypertable ... New 'Analytics' emerge MapReduce, Hadoop ... New 'Compute' New 'Data'
17. How do the new guys compete? Web server app server database server Storage system Web server Web server app server database server Storage system Load balancer distributed memory cache
24. Automation – scaling from 50 to 3500 servers in 3 days 4/11 4/12 4/13 4/13 4/14 4/15 4/15 4/16 4/17 4/17 4/18 4000 3000 2000 1500 1000 500 50 Animoto case study Servers
25. Map Reduce design pattern Documents Map <word, cnt> Map <word, cnt> Map <word, cnt> hash(word, mod R) Reduce <word, cnt> Reduce <word, cnt> Reduce <word, cnt> Result <apple, 2,045,455> <barn, 254,345> Example: compute word frequency over a large set of documents Map <word, cnt> Result <house, 111,341> <kitchen, 4,678> Result <cat, 1,033,746> <horse, 25,387> Reduce workers: reads and sorts Map output, invoking Reduce() task which sums counts for each word Map workers: reads input, invoking Map() function which finds words and outputs records for reduce workers P 1 P 1 P 1 P 1