These Five Tricks Can Make Your Apps Greener, Cheaper, & Nicer

The code we write has a climate impact. But how big is that impact? How do we measure it? How do we reduce it? Is the cloud helping? What’s going on with Virginia? Are we still allowed to do CI/CD? Will native compilation save us? Is Java even a good choice anymore? This talk discusses some of the trade-offs for a modern software developer, and provides a roadmap to figuring out the right thing.

Disclosure: Holly works on Quarkus. Along the way, she will talk about Quarkus sustainability measurements we’ve been doing … but it’s ok, because she promises that the Quarkus carbon data is exciting and interesting.

Transcript

1. Holly Cummins Senior Principal Software Engineer, Quarkus GOTO Aarhus May 24, 2023 @holly_cummins@hachyderm.io These five tricks can make your apps greener cheaper & nicer
2. @holly_cummins #RedHat
3. @holly_cummins #RedHat compute for training large deep learning models 300,000-fold increase in 6 years Dario Amodei and Daniel Hernandez. 2018. AI and Compute. https://openai.com/blog/ai-and-compute
4. @holly_cummins #RedHat compute for training large deep learning models 300,000-fold increase in 6 years Dario Amodei and Daniel Hernandez. 2018. AI and Compute. https://openai.com/blog/ai-and-compute
5. @holly_cummins #RedHat it’s not just artificial intelligence
6. @holly_cummins #RedHat it’s not just artificial intelligence it’s not just cryptocurrency mining
7. @holly_cummins #RedHat it’s not just artificial intelligence it’s not just cryptocurrency mining it’s all of us
8. @holly_cummins #RedHat Sources: https://www.greenit.fr/wp-content/uploads/2019/11/GREENIT_EENM_etude_EN_accessible.pdf https://ourworldindata.org/ghg-emissions-by-sector the digital world creates more carbon emissions than aviation
9. @holly_cummins #RedHat Lighter area represents high and low estimates, where available. Sources: https://www.iea.org/fuels-and-technologies/data-centres-networks https://ourworldindata.org/grapher/electricity-demand?tab=table&country=USA~GBR~FRA~DEU~IND~BRA data centres use as much electricity as a medium country
10. @holly_cummins #RedHat aaaaaaaargh
11. @holly_cummins #RedHat aaaaaaaargh?
12. @holly_cummins #RedHat be a solutionist
13. @holly_cummins #RedHat how do we do make solutions?
14. @holly_cummins #RedHat green software foundation: principles
15. @holly_cummins #RedHat carbon awareness green software foundation: principles
16. @holly_cummins #RedHat carbon awareness green software foundation: principles where
17. @holly_cummins #RedHat carbon awareness green software foundation: principles where when
18. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency where when
19. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency where when elasticity
20. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency where when elasticity utilisation
21. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency electricity efficiency where when elasticity utilisation
22. @holly_cummins #RedHat algorithms carbon awareness green software foundation: principles hardware efficiency electricity efficiency where when elasticity utilisation
23. @holly_cummins #RedHat algorithms stack carbon awareness green software foundation: principles hardware efficiency electricity efficiency where when elasticity utilisation
24. @holly_cummins #RedHat algorithms stack carbon awareness hardware efficiency electricity efficiency where when elasticity utilisation green software foundation: principles
25. trick 1: electricity source
26. @holly_cummins #RedHat data center location matters
27. @holly_cummins #RedHat
28. @holly_cummins #RedHat
29. @holly_cummins #RedHat we need to talk about virginia
30. @holly_cummins #RedHat look at the sustainability information before choosing a hosting region
31. @holly_cummins #RedHat look at the sustainability information before choosing a hosting region choose a cloud provider who make this easy
32. @holly_cummins #RedHat time of day matters • (most) renewables are intermittent • if grid load is high, shortfalls are filled by fossil fuels
33. @holly_cummins #RedHat move the workload to the greenest place
34. @holly_cummins #RedHat move the workload to the greenest place
35. @holly_cummins #RedHat move the workload to the greenest place a data problem
36. @holly_cummins #RedHat move the workload to the greenest place a data problem an orchestration problem
37. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency electricity efficiency
38. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency electricity efficiency elasticity
39. @holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency electricity efficiency elasticity utilisation
40. #RedHat @holly_cummins 2017 server-survey 25% doing no useful work (16,000 sampled)
41. #RedHat @holly_cummins 2017 server-survey 25% doing no useful work (16,000 sampled) “perhaps someone forgot to turn them off”
42. #RedHat @holly_cummins@hachyderm.io https://www.anthesisgroup.com/wp-content/uploads/2019/11/Comatose-Servers-Redux-2017.pdf 2014 server-survey 29% active less than 5% of the time (4,000 sampled)
43. @holly_cummins #RedHat
44. @holly_cummins #RedHat cloud elasticity? https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033
45. @holly_cummins #RedHat cloud elasticity? 2021: https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033
46. @holly_cummins #RedHat cloud elasticity? 2021: $26.6 billion wasted https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033
47. @holly_cummins #RedHat cloud elasticity? 2021: $26.6 billion wasted by always-on cloud instances https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033
48. @holly_cummins #RedHat elasticity we used to leave our applications running all the time when we scripted turning them off at night, we reduced our cloud bill by 30% @darkandnerdy, Chicago DevOpsDays
49. @holly_cummins #RedHat it’s not just electricity
50. @holly_cummins #RedHat it’s not just electricity it’s water
51. @holly_cummins #RedHat it’s not just electricity it’s water it’s e-waste
52. @holly_cummins #RedHat yes, turning applications off is scary
53. @holly_cummins #RedHat what if … turning applications off was no more scary than turning the lights off?
54. @holly_cummins #RedHat ultimate elasticity
55. @holly_cummins #RedHat ultimate elasticity
56. @holly_cummins #RedHat turning it off and on again must ultimate elasticity
57. @holly_cummins #RedHat turning it off and on again must • be fast ultimate elasticity
58. @holly_cummins #RedHat turning it off and on again must • be fast • actually work ultimate elasticity
59. @holly_cummins #RedHat turning it off and on again must • be fast • actually work • idempotency ultimate elasticity
60. @holly_cummins #RedHat turning it off and on again must • be fast • actually work • idempotency • resiliency ultimate elasticity
61. @holly_cummins #RedHat architect things to be turned off and on often
62. @holly_cummins #RedHat trick 2: LightSwitchOps architect things to be turned off and on often
63. @holly_cummins #RedHat trick 2: LightSwitchOps architect things to be turned off and on often
64. #Quarkus @holly_cummins you can’t optimise what you can’t measure
65. @holly_cummins #RedHat FinOps figuring out who in your company forgot to turn off their cloud
66. @holly_cummins #RedHat
67. @holly_cummins #RedHat backstage.io
68. @holly_cummins #RedHat backstage.io •cost insights plugin
69. @holly_cummins #RedHat backstage.io •cost insights plugin •cloud carbon footprint plugin
70. @holly_cummins #RedHat algorithms stack carbon awareness hardware efficiency electricity efficiency where when elasticity utilisation green software foundation: principles
71. @holly_cummins #RedHat efficiency
72. @holly_cummins #RedHat what programming languages use the least energy?
73. @holly_cummins #RedHat what programming languages use the most energy?
75. energy efficiency of programming languages
76. @holly_cummins #RedHat
77. @holly_cummins #RedHat
78. @holly_cummins #RedHat what’s the most carbon-efficient java?
79. @holly_cummins #RedHat
80. @holly_cummins #RedHat
81. @holly_cummins #RedHat does being small and fast reduce carbon footprint?
82. #Quarkus @holly_cummins measure, don’t guess.
83. digression: measuring carbon is hard
84. @holly_cummins #RedHat step 1: measure power usage
85. @holly_cummins #RedHat step 1: measure power usage wall power measurement
86. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete
87. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete needs access to the wall
88. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete needs access to the wall and equipment
89. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete needs access to the wall and equipment RAPL
90. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete needs access to the wall and equipment RAPL programmatically accessible
91. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete needs access to the wall and equipment RAPL programmatically accessible misses some components
92. @holly_cummins #RedHat step 1: measure power usage wall power measurement more complete needs access to the wall and equipment data costs carbon RAPL programmatically accessible misses some components
93. @holly_cummins #RedHat
94. @holly_cummins #RedHat
95. @holly_cummins #RedHat
96. @holly_cummins #RedHat load
97. @holly_cummins #RedHat Source: Teads EC2 instances carbon dataset
98. @holly_cummins #RedHat coal wind step 2: convert power usage to carbon solar
99. @holly_cummins #RedHat published energy mixes (these are made-up energy mixes)
100. @holly_cummins #RedHat published energy mixes … but methodologies are not open (these are made-up energy mixes)
101. @holly_cummins #RedHat published energy mixes … but methodologies are not open … or consistent (these are made-up energy mixes)
102. @holly_cummins #RedHat step 3: embedded carbon (manufacturing has costs)
103. @holly_cummins #RedHat simpler models
104. @holly_cummins #RedHat all models are wrong, some are useful
105. @holly_cummins #RedHat trick 3: vrrrrrooooooooom model* * a made-up name
106. @holly_cummins #RedHat
107. @holly_cummins #RedHat these two columns are almost the same
108. @holly_cummins #RedHat energy consumption (sort of, mostly) is proportional to execution time
109. @holly_cummins #RedHat trick 4: economic model*
110. @holly_cummins #RedHat trick 4: economic model* * "economic input-output life cycle assessment"
111. @holly_cummins #RedHat reducing your cloud spend (probably) reducing your carbon footprint* hardware spend electricity bill * if you keep other factors the same
112. @holly_cummins #RedHat worked(ish) example: what’s the carbon footprint of ChatGPT?
113. @holly_cummins #RedHat worked(ish) example: what’s the carbon footprint of ChatGPT?
114. @holly_cummins #RedHat worked(ish) example: what’s the carbon footprint of ChatGPT? $50,000- $700,000 running costs per day https://www.businessinsider.com/how-much-chatgpt-costs-openai-to-run-estimate-report-2023-4
115. @holly_cummins #RedHat worked(ish) example: what’s the carbon footprint of ChatGPT? $3,000,000- $50,000,000 training costs $50,000- $700,000 running costs per day https://www.businessinsider.com/how-much-chatgpt-costs-openai-to-run-estimate-report-2023-4
116. so … quarkus?
117. @holly_cummins #RedHat density Source: Clement Escoffier cost impact of framework choice Setup: • 800 requests/second, over 20 days • SLA > 99% • AWS instances Assumptions: • Costs are for us-east-1 data centre
118. @holly_cummins #RedHat Setup: • 800 requests/second, over 20 days • SLA > 99% Assumptions: • 50% load • us-east-1 data centre • Teads dataset Source: Clement Escoffier x Teads cloud carbon impact of framework choice
119. @holly_cummins #RedHat Setup: • 800 requests/second, over 20 days • SLA > 99% Assumptions: • 50% load • us-east-1 data centre • Teads dataset Source: Clement Escoffier x Teads cloud carbon impact of framework choice economic model in action: the cost and carbon metrics are (roughly) the same
120. @holly_cummins #RedHat Setup: • REST + CRUD • large heap • RAPL energy measurement • multiple instances to support high load 
 Assumptions: • US energy mix Source: John O’Hara climate impact as a function of load
121. @holly_cummins #RedHat Setup: • REST + CRUD • large heap • RAPL energy measurement • multiple instances to support high load 
 Assumptions: • US energy mix Source: John O’Hara climate impact as a function of load vrrrooom model in action: quarkus on JVM has the smallest footprint … because it has the highest throughput
122. @holly_cummins #RedHat what about memory? more complete model: both throughput (execution time) and memory contribute to carbon footprint
123. @holly_cummins #RedHat what about memory? more complete model: both throughput (execution time) and memory contribute to carbon footprint quarkus is most efficient for • startup time • throughput • memory (RSS size)
124. @holly_cummins #RedHat trick 5: use quarkus
125. @holly_cummins #RedHat trick 5: use quarkus quarkus ‘automatically’ saves
126. @holly_cummins #RedHat trick 5: use quarkus quarkus ‘automatically’ saves • time
127. @holly_cummins #RedHat trick 5: use quarkus quarkus ‘automatically’ saves • time • money
128. @holly_cummins #RedHat trick 5: use quarkus quarkus ‘automatically’ saves • time • money • carbon (~2x)
129. @holly_cummins #RedHat trick 5: use quarkus quarkus ‘automatically’ saves • time • money • carbon (~2x) • … even when Spring compatibility libraries are used (almost no code changes except dependencies and tests)
130. @holly_cummins #RedHat trick 5: use quarkus quarkus ‘automatically’ saves • time • money • carbon (~2x) • … even when Spring compatibility libraries are used (almost no code changes except dependencies and tests)
131. @holly_cummins #RedHat aaaaaaaargh?
132. @holly_cummins #RedHat carbon awareness hardware efficiency electricity efficiency green software foundation: principles utility
133. @holly_cummins #RedHat is using this carbon giving value? utility
134. @holly_cummins #RedHat “no-regrets” solutions
135. @holly_cummins #RedHat
136. @holly_cummins #RedHat co-benefits
137. @holly_cummins #RedHat co-benefits the double win
138. @holly_cummins #RedHat co-benefits the double win win-win
139. @holly_cummins #RedHat co-benefits the double win win-win 1 + 1 = 3
140. @holly_cummins #RedHat co-benefits the double win win-win 1 + 1 = 3 twofer
141. @holly_cummins #RedHat co-benefits the double win win-win 1 + 1 = 3 twofer überwinden
142. @holly_cummins #RedHat co-benefits the double win win-win 1 + 1 = 3 twofer überwinden
143. @holly_cummins #RedHat climate solutions can make everything better
144. @holly_cummins #RedHat remember the zombie servers? https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033
145. @holly_cummins #RedHat remember the zombie servers? what else could we do with that $26.6 billion of wasted cloud spend? https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033
146. @holly_cummins #RedHat the double-win turning things off saves a lot of money
147. @holly_cummins #RedHat the double-win renewable electricity is 9x cheaper
148. @holly_cummins #RedHat the double-win renewable electricity is 9x cheaper hosting in Montreal:
149. @holly_cummins #RedHat the double-win renewable electricity is 9x cheaper hosting in Montreal: 88% less carbon than the same workload in London
150. @holly_cummins #RedHat the double-win renewable electricity is 9x cheaper hosting in Montreal: 88% less carbon than the same workload in London and it’s 15% cheaper
151. @holly_cummins #RedHat remember the vrrrrrooooooooom model? (probably not, it was a made-up name)
152. @holly_cummins #RedHat we need a new mental model for ‘virtuous’ car: high max speed means high fuel usage per mile travelled
153. @holly_cummins #RedHat software: high max transactions means low carbon per transaction we need a new mental model for ‘virtuous’ car: high max speed means high fuel usage per mile travelled
154. @holly_cummins #RedHat the double-win
155. @holly_cummins #RedHat the double-win
156. @holly_cummins #RedHat “this is not sacrifice. it’s advancement.” – Dr. Jonathan Foley
157. @holly_cummins #RedHat tl;dpa (too long; didn’t pay attention) trick 1: choose your hosting wisely trick 2: architect to be able to turn stuff off (LightSwitchOps) trick 3: the vrrrooooom model says faster is greener trick 4: the economic model says cheaper is greener trick 5: choose a fast and energy-efficient framework, such as quarkus
158. @holly_cummins #RedHat we all make a difference
160. slides thank you @holly_cummins@hachyderm.io