Writing Greener Java Applications

Holly Cummins Senior Principal Software Engineer, Quarkus Devoxx UK May
10, 2023 @[email protected] Writing Greener Java Applications

@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

@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

@holly_cummins #RedHat aaaaaaaargh

@holly_cummins #RedHat aaaaaaaargh?

@holly_cummins #RedHat be a solutionist

@holly_cummins #RedHat what should we do?

@holly_cummins #RedHat should we reduce our greenhouse gases?

@holly_cummins #RedHat • right thing to do should we reduce
our greenhouse gases?

@holly_cummins #RedHat • right thing to do • business benefits
should we reduce our greenhouse gases?

• lower costs should we reduce our greenhouse gases?

• lower costs • higher ESG score should we reduce our greenhouse gases?

• lower costs • higher ESG score • competitive advantage should we reduce our greenhouse gases?

• lower costs • higher ESG score • competitive advantage • recruitment advantage should we reduce our greenhouse gases?

@holly_cummins #RedHat green software foundation: principles

@holly_cummins #RedHat carbon awareness green software foundation: principles

@holly_cummins #RedHat carbon awareness green software foundation: principles where

@holly_cummins #RedHat carbon awareness green software foundation: principles where when

@holly_cummins #RedHat carbon awareness green software foundation: principles hardware efficiency
where when

where when elasticity

where when elasticity utilisation

electricity efficiency where when elasticity utilisation

@holly_cummins #RedHat algorithms carbon awareness green software foundation: principles hardware
efficiency electricity efficiency where when elasticity utilisation

@holly_cummins #RedHat algorithms stack carbon awareness green software foundation: principles
hardware efficiency electricity efficiency where when elasticity utilisation

@holly_cummins #RedHat algorithms stack carbon awareness hardware efficiency electricity efficiency
where when elasticity utilisation green software foundation: principles

step 1: electricity source

@holly_cummins #RedHat data center location matters

@holly_cummins #RedHat

@holly_cummins #RedHat we need to talk about virginia

@holly_cummins #RedHat look at the sustainability information before choosing a
hosting region

@holly_cummins #RedHat look at the sustainability information before choosing a
hosting region choose a cloud provider who make this easy

@holly_cummins #RedHat time of day matters • (most) renewables are
intermittent • if grid load is high, shortfalls are filled by fossil fuels

@holly_cummins #RedHat move the workload to the greenest place

@holly_cummins #RedHat move the workload to the greenest place a
data problem

@holly_cummins #RedHat move the workload to the greenest place a
data problem an orchestration problem

electricity efficiency

electricity efficiency elasticity

electricity efficiency elasticity utilisation

@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

#RedHat @holly_cummins 2017 survey 25% of 16,000 servers doing no
useful work

#RedHat @holly_cummins 2017 survey 25% of 16,000 servers doing no
useful work “perhaps someone forgot to turn them off”

#RedHat @[email protected] 2014 survey 29% of 4,000 active less than
5% of the time https://www.anthesisgroup.com/wp-content/uploads/2019/11/Comatose-Servers-Redux-2017.pdf

@holly_cummins #RedHat yes, turning applications off is scary

@holly_cummins #RedHat ultimate elasticity

@holly_cummins #RedHat turning it off and on again must ultimate
elasticity

@holly_cummins #RedHat turning it off and on again must •
be fast ultimate elasticity

be fast • actually work ultimate elasticity

be fast • actually work • idempotency ultimate elasticity

be fast • actually work • idempotency • resiliency ultimate elasticity

@holly_cummins #RedHat LightSwitchOps

@holly_cummins #RedHat algorithms stack carbon awareness hardware efficiency electricity efficiency
where when elasticity utilisation green software foundation: principles

@holly_cummins #RedHat efficiency

@holly_cummins #RedHat what programming languages use the least energy?

@holly_cummins #RedHat what programming languages use the most energy?

energy efficiency of programming languages

@holly_cummins #RedHat “Our results show that some JVM platforms can
exhibit up to 100% more energy consumption.” https://hal.inria.fr/hal-03275286/document

@holly_cummins #RedHat if JVM parameters can make that much difference
… … what impact does the framework have?

@holly_cummins #RedHat does being small and fast reduce carbon footprint?

#Quarkus @holly_cummins measure, don’t guess.

digression: measuring carbon is hard

@holly_cummins #RedHat step 1: measure power usage

@holly_cummins #RedHat step 1: measure power usage wall power measurement

more complete

more complete needs access to the wall

more complete needs access to the wall and equipment

more complete needs access to the wall and equipment and actual paper

more complete needs access to the wall and equipment and actual paper RAPL

more complete needs access to the wall and equipment and actual paper RAPL programmatically accessible

more complete needs access to the wall and equipment and actual paper RAPL programmatically accessible misses some components

more complete needs access to the wall and equipment and actual paper data costs carbon RAPL programmatically accessible misses some components

@holly_cummins #RedHat load

@holly_cummins #RedHat Source: Teads EC2 instances carbon dataset

@holly_cummins #RedHat coal wind step 2: convert power usage to
carbon solar

@holly_cummins #RedHat published energy mixes (these are made-up energy mixes)

@holly_cummins #RedHat published energy mixes … but methodologies are not
open (these are made-up energy mixes)

@holly_cummins #RedHat published energy mixes … but methodologies are not
open … or consistent (these are made-up energy mixes)

@holly_cummins #RedHat step 3: embedded carbon (manufacturing has costs)

@holly_cummins #RedHat simpler models

@holly_cummins #RedHat vrrrrrooooooooom model* * a made-up name

@holly_cummins #RedHat these two columns are almost the same

@holly_cummins #RedHat energy consumption (sort of, mostly) is proportional to
execution time

@holly_cummins #RedHat economic model*

@holly_cummins #RedHat economic model* * "economic input-output life cycle assessment"

@holly_cummins #RedHat reducing your cloud spend (probably) reducing your carbon
footprint* hardware spend electricity bill * if you keep other factors the same

@holly_cummins #RedHat all models are wrong, some are useful

so … quarkus?

@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

@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

• quarkus cuts carbon by ~2x* • native has a
slightly smaller carbon footprint * OOM heuristic density test: conclusions

@holly_cummins #RedHat capacity Source: John O’Hara Setup: • REST +
CRUD • large heap • RAPL energy measurement Assumptions: • US energy mix climate impact at low load (single instance)

@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 at high load (multiple instances)

• quarkus cuts carbon by ~2-3x* • native consumes more
carbon than JVM capacity test: conclusions * OOM heuristic

@holly_cummins #RedHat so does quarkus reduce carbon emissions?

@holly_cummins #RedHat so does quarkus reduce carbon emissions? quarkus ‘automatically’
saves

saves • time

saves • time • money

saves • time • money • carbon (~2x)

@holly_cummins #RedHat but what about native? reminder

@holly_cummins #RedHat but what about native? use case for native
reminder

• low workload (throughput isn’t the bottleneck) reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) • serverless (of course) reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) • serverless (of course) use case for JVM reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) • serverless (of course) use case for JVM • high workload (you need lots of throughput) reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) • serverless (of course) use case for JVM • high workload (you need lots of throughput) • long-lived processes (the rapid start of native doesn’t save you much over the lifetime) reminder

• low workload (throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) • serverless (of course) use case for JVM • high workload (you need lots of throughput) • long-lived processes (the rapid start of native doesn’t save you much over the lifetime) • stable workload or very little elasticity in underlying orchestration reminder

@holly_cummins #RedHat use case for native • low workload (so
throughput isn’t the bottleneck) • resource-constrained or old hardware (especially memory) • high re-deploy rate (applications never get warmed up before being spun down) • serverless (of course)   use case for JVM • high workload (you need lots of throughput) • long-lived processes (the rapid start of native doesn’t save you much over the lifetime) • stable workload or very little elasticity in underlying orchestration but what about native? to optimise carbon

@holly_cummins #RedHat yes, those are the same optimising • cost
• application performance will also be optimising carbon

@holly_cummins #RedHat native vs JVM is not an either-or •
consider hybrid deployment models • a strength of Quarkus is application code is the same for native and JVM

@holly_cummins #RedHat native CLIs? • native is not just for
serverless • native Java makes great CLIs • is a CLI usually the most carbon- efficient way to do something?

@holly_cummins #RedHat elasticity native quarkus starts faster than a light
bulb

@holly_cummins #RedHat but. an app starting fast isn’t elasticity …
…unless you also stop the app

@holly_cummins #RedHat how does quarkus help reduce carbon? energy usage
elasticity machine selection (provisioning)

‘free’ direct easy to measure elasticity machine selection (provisioning)

‘free’ direct easy to measure lower memory usage higher throughput for same resources elasticity machine selection (provisioning)

‘free’ direct easy to measure lower memory usage higher throughput for same resources ‘enabled behaviours’ harder to measure real world impact depends on people taking advantage elasticity machine selection (provisioning)

‘free’ direct easy to measure lower memory usage higher throughput for same resources ‘enabled behaviours’ harder to measure real world impact depends on people taking advantage elasticity machine selection (provisioning) running the same workload on a smaller machine saves energy saves embodied carbon

‘free’ direct easy to measure lower memory usage higher throughput for same resources ‘enabled behaviours’ harder to measure real world impact depends on people taking advantage elasticity machine selection (provisioning) scaling workloads down (ideally to 0) serverless a good example (but not the only one) running the same workload on a smaller machine saves energy saves embodied carbon

@holly_cummins #RedHat aaaaaaaargh?

@holly_cummins #RedHat remember the vrrrrrooooooooom model? (probably not, it was
a made-up name)

@holly_cummins #RedHat we need a new mental model car: high
max speed means high fuel usage per mile travelled

@holly_cummins #RedHat software: high max transactions means low carbon per
transaction we need a new mental model car: high max speed means high fuel usage per mile travelled

@holly_cummins #RedHat carbon awareness hardware efficiency electricity efficiency green software
foundation: principles utility

@holly_cummins #RedHat is using this carbon giving value? utility

@holly_cummins #RedHat “no-regrets” solutions

@holly_cummins #RedHat co-benefits

@holly_cummins #RedHat co-benefits the double win

@holly_cummins #RedHat co-benefits the double win win-win

@holly_cummins #RedHat co-benefits the double win win-win win-squared

@holly_cummins #RedHat co-benefits the double win win-win win-squared 1 +
1 = 3

1 = 3 twofer

1 = 3 twofer überwinden

1 = 3 twofer überwinden the extra win

@holly_cummins #RedHat climate solutions can make everything better

@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

@holly_cummins #RedHat remember the zombie servers? $26.6 billion https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033

@holly_cummins #RedHat remember the zombie servers? $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

@holly_cummins #RedHat remember the zombie servers? $26.6 billion wasted in
2021 https://www.business2community.com/cloud-computing/overprovisioning-always-on-resources-lead-to-26-6-billion-in-public-cloud-waste-expected-in-2021-02381033

@holly_cummins #RedHat the double-win turning things off saves a lot
of money

@holly_cummins #RedHat the double-win renewable electricity is 9x cheaper

@holly_cummins #RedHat the double-win renewable electricity is 9x cheaper hosting
in Montreal:

in Montreal: 88% less carbon than the same workload in London

in Montreal: 88% less carbon than the same workload in London and it’s 15% cheaper

@holly_cummins #RedHat the double-win

@holly_cummins #RedHat “this is not sacrifice. it’s advancement.”

@holly_cummins #RedHat we all have a part to play •
choose your hosting wisely • use your buying power to encourage businesses to do make greener choices • turn things off when you’re not using them • choose an energy-efficient framework • all of this is a win anyway

slides thank you @[email protected]

Writing Greener Java Applications

Writing Greener Java Applications

More Decks by Holly Cummins

Other Decks in Programming

Featured

Transcript