FrOSCon 2017 talk about vutuv

The talk was about some of the business and technical challenges (e.g. WebPerformance) creating and hosting https://www.vutuv.de . https://programm.froscon.de/2017/events/1960.html

Transcript

1. fast, free and secure business networking https://www.vutuv.de
2. A Business Network Think LinkedIn or XING but open-source (MIT). Free, fast and secure.
3. Why would one need an other business network?
4. Nobody likes LinkedIn or XING. Too many annoying emails and pop-ups. Nobody likes to pay for premium accounts. But those are First World problems!
5. LinkedIn and XING only work well 
 in a few countries (but still slow).
6. The majority of potential business networking users use low bandwidth feature phones which don’t support JavaScript. They have no access to LinkedIn at all. 
 And they can’t pay for a premium account.
7. Some users need extra security. Not just HTTPS but Tor. Think oppressive governments and companies which use proxies to sniff if their employees hunt for new jobs.
8. Unique Selling Points • vutuv accounts are free.
 Revenue is generated with ads and human resources features. • vutuv is available.
 LinkedIn can’t be used in most parts of the world. vutuv is part of internet.org and can cover the whole world. • vutuv is fast.
 3-8 times faster than LinkedIn. • vutuv is secure.
 Your boss or government can’t see if you are hunting for a new job. Not just https but Tor.
 Bonus: We don’t store passwords. We don’t use cookies for non logged in users. • vutuv profiles have a better Google ranking
 Because of the side’s speed an average profile gets a better Google ranking.
 For many users it’s there first business networking account at all which results in unique content.
9. Traction Number of accounts 0 250.000 500.000 750.000 1.000.000 Nov. 2016 Dec. 2016 Jan. 2017 Feb. 2017 Mar. 2017 Apr. 2017 April: > 5.000.000 unique visitors per month
10. > 1,000,000 users?! WOW! Did you get VC money?
 Did you buy a new car?
11. No, I did not get any VC money.
 
 By now I have given up on it and don’t search for investors any more.
12. Technology Today • Phoenix Framework
 Runs on the Erlang-BEAM like WhatsApp. • MySQL • No JavaScript
 Runs on all devises! Source-Code: https://github.com/vutuv/vutuv/
13. vutuv differentiates itself from other business networks by speed.
14. WebPerformance Fully Loaded on the browser: 700 ms (DSL in Germany)
15. Loading Time from Dulles, VA https://www.webpagetest.org/video/compare.php?tests=170131_QQ_BHG,170131_9F_BHH,170131_49_BHJ DSL
16. Why is WebPerformance important?
17. Jakob Nielsen’s book „Usability Engineering“ from 1993 Delay User reaction 0 - 100 ms Instant 100 - 300 ms Feels sluggish 300 - 1000 ms Machine is working… 1 s+ Mental context switch 10 s+ I’ll come back later…
18. Web Search Delay Experiment Type of Delay Delay (ms) Duration (weeks) Impact on Avg. Daily Searches Pre-header 100 4 -0.20 % Pre-header 200 6 -0.59% Post-header 400 6 0.59% Post-ads 200 4 0.30% Source: https://www.igvita.com/slides/2012/webperf-crash-course.pdf
19. Numbers by SOASTA
20. Numbers by SOASTA
21. Network latency is our biggest problem. Why?
22. TCP Source: http://en.wikipedia.org/wiki/Transmission_Control_Protocol Latency
23. Download a 58 kB file from a US east cost server to a client in Frankfurt. Protocol: HTTP Roundtrip Time: 80 ms Zeit client server 0 ms SYN --> <-- SYN,ACK 80 ms ACK --> GET --> <-- 10 TCP Segmente (14.600 Bytes) 160 ms ACK --> <-- 20 TCP Segmente (29.200 Bytes) 240 ms ACK --> <-- 40 TCP Segmente (15.592 Bytes) 320 ms ACK --> Time to download a 58 kB file: 320 ms
24. TCP Slow-Start KB 0 55 110 165 220 Roundtrip 1. 2. 3. 4. 213 99 42 14 114 57 28 14
25. SSL adds at least 
 one roundtrip.
 
 HTTP/2.0 needs SSL.
26. HTTPS KB 0 50 100 150 200 Roundtrip 1. 2. 3. 4. 199 85 28 0 114 57 28 14 SSL
27. Above the fold target: 28 KB KB 0 50 100 150 200 Roundtrip 1. 2. 3. 4. 199 85 28 0 114 57 28 14 SSL
28. Latency is king! Bandwidth is not important for normal websites.
29. Frankfurt, Germany Sydney, Australia Sydney: https://www.webpagetest.org/result/170130_3H_S13/
 Frankfurt: https://www.webpagetest.org/result/170130_YJ_SRS/ DSL
30. Frankfurt, Germany Sydney, Australia Sydney: https://www.webpagetest.org/result/170130_3H_S13/
 Frankfurt: https://www.webpagetest.org/result/170130_YJ_SRS/ DSL
31. Frankfurt: https://www.webpagetest.org/result/170131_CX_B7D/
 Sydney: https://www.webpagetest.org/result/170131_XZ_B7P/ 3G Frankfurt, Germany Sydney, Australia
32. Frankfurt: https://www.webpagetest.org/result/170131_CX_B7D/
 Sydney: https://www.webpagetest.org/result/170131_XZ_B7P/ 3G Frankfurt, Germany Sydney, Australia
33. How do we tackle WebPerformance?
34. Set a time budget and stick to it! Our time budget is 1 second within Germany.
35. Start Render Document Complete
36. Important: We can not control the network!
37. What we can control: • Transfer Protocol (e.g. HTTP/1.1 vs. HTTP/2.0) • Compression (e.g. gzip vs. zopfli vs. brotli) • Amount of files • File size • Time the server needs to generate the HTML • Content (e.g. HTML, CSS, JavaScript, Images)
39. File size Time on the server
40. The Servers
41. We use NGINX, MySQL and Phoenix Framework. Bare metal. Debian Linux.
42. Phoenix is about 
 10 times faster 
 than Ruby on Rails. Yes, I did a proof of concept of vutuv with Ruby on Rails.
43. But just a fast programming language is not good enough! Remember: We have to run vutuv.de on a dime.
44. We avoid serving freshly generated HTML when possible. Hard Drives are much cheaper than CPU.
45. NGINX is our gatekeeper. It checks for cookies and routes accordingly. I did think of Varnish but hard drive is cheaper than RAM.
46. /var/www/vutuv/ cache - index.html - index.html.gz users - xyz.html - xyz.html.gz No Cookie = Static Content Cross-Site Request Forgery (CSRF)
47. 1,000,000 users x 28 KB = 26 GB
48. Rendering a user’s profile: 50 ms. Serving a static file: less than 1 ms.
49. The Avatars A little bit more than just https://github.com/stavro/arc
56. guetzli -quality 75 a.jpg /tmp/q75-a.jpg convert a.jpg /tmp/q75-a.jpg -fx 'hypot(65-i, 65-j) < 65 ? u : v' new_a.jpg https://github.com/google/guetzli
57. By that we save 15-20% of file size. We use https://github.com/stavro/arc for uploading and initial conversion and optimize during off peak times.
58. Compression
59. /var/www/vutuv/ cache - index.html - index.html.gz users - xyz.html - xyz.html.gz Provide a compressed version of a static file.
60. html-minifier --case-sensitive --collapse-boolean-attributes --collapse-whitespace --remove-comments --remove-optional-tags --remove-tag-whitespace --sort-attributes --sort-class-name --collapse-inline-tag-whitespace --conservative-collapse a.html > b.html Optimize before compress
61. Use zopfli instead of gzip. https://github.com/google/zopfli
62. For Phoenix live content use brotli in NGINX. https://github.com/google/brotli
63. Inlining
64. HTTP caching is good but for our time budget not good enough.
65. We inline a lot.
66. We inline all the CSS. Obviously we don’t use Twitter Bootstrap.
67. We inline many images.
70. Fill the Steps KB 0 50 100 150 200 Roundtrip 1. 2. 3. 4. 199 85 28 0 114 57 28 14 SSL
71. Prefetch
72. Analyse your workflow and prefetch stuff that you’ll need in the next step. <head> […] <%= if @conn.assigns[:prefetch] do %> <link rel="prefetch" title="a" href="a.jpg"> <% end %> </head>
74. The future of vutuv?
75. Obviously money is the biggest problem. Because we don’t use JavaScript we can’t use the classic Google Ads system to earn money. And charging for accounts is a no go either.
76. Potential Future Technology • Phoenix Framework
 Runs on the Erlang-BEAM like WhatsApp. • CouchDB
 Syncs over continents. Super fast anywhere. • No JavaScript
 Runs on all devises! • Cheap servers
 The world is our data center (blue dots). We are not affected by catastrophes. DNS will reroute to other continents. • Short distance to the browser
 This ensures the best browsing experience.
77. You can help! Please create an account and tell your friends about www.vutuv.de
78. www.wintermeyer-consulting.de Ruby on Rails, Phoenix und WebPerformance. twitter.com/wintermeyer