Thinking distributed

Thinking distributed about your approach to building systems and how node.js is ideal for building systems in a distributed way.

Transcript

1. paolofragomeni github.com/hij1nx co-founder of nodejitsu and the node firm twitter.com/hij1nx Saturday, September 29, 12
2. is not easy distributed thinking 1 Saturday, September 29, 12
3. 2 “scale” more vertical more memory more power more cpu Saturday, September 29, 12
4. 3 wtf Saturday, September 29, 12
5. 4 plateau monolithic code A Saturday, September 29, 12
6. 4 plateau monolithic code A B cognitive indigestion Saturday, September 29, 12
7. 4 plateau monolithic code A C responsibility overload B cognitive indigestion Saturday, September 29, 12
8. 4 plateau monolithic code A D individual failures become system failures C responsibility overload B cognitive indigestion Saturday, September 29, 12
9. complexity 5 quantifiable as the number of steps required to bring a system to its intended state. Saturday, September 29, 12
10. FAILURE 6 Saturday, September 29, 12
11. api endpoint img processing email processing auth 7 separation of concerns img email auth client client api Saturday, September 29, 12
12. 8 tradition service oriented clients usually fail and the process ends when the endpoint is unreachable. Saturday, September 29, 12
13. a case study jitsu is a deployment tool written in javascript 9 Saturday, September 29, 12
14. another case study github is a hosting solution for git projects. wut? 10 “ssh: Could not resolve hostname github.com: nodename nor servname provided, or not known fatal: The remote end hung up unexpectedly” Saturday, September 29, 12
15. a solution d is a generic deployment tool and the basis for the next version of jitsu. 11 Saturday, September 29, 12
16. distribution of 12 api endpoint img processing email processing auth img email auth client client api responsibility Saturday, September 29, 12
17. distribution of 12 api endpoint img processing email processing auth img email auth client client api api api responsibility Saturday, September 29, 12
18. a demonstration 13 Saturday, September 29, 12
19. conclusion 14 in a network context, centralized systems are less reliable. A Saturday, September 29, 12
20. conclusion 14 in a network context, centralized systems are less reliable. A the number of end points should correlate to the demand on the service. B Saturday, September 29, 12
21. conclusion 14 in a network context, centralized systems are less reliable. A the number of end points should correlate to the demand on the service. B no single end point should be more important than another. C Saturday, September 29, 12
22. thanks X Saturday, September 29, 12