An innocent's guide to ignoring best practices

AWS User Group Bengaluru
AWS User Group BengaluruAWS User Group Bengaluru
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
BENGALURU
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
An Innocent’s Guide To
Ignoring Best
Practices...
AMIT GARDE, VP Engineering, Swiggy | 6 October 2018
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Preliminaries
• What this talk is *not* going to be…
• Prescriptive
• A laundry list of things to do or to not do
• What this talk is *trying* to be
• A perspective on how “Best Practice” emerges, why a grain of salt is useful, and
why we could usefully borrow ideas from other fields of human endeavour
• What this talk *might* end up being
• Whimsical
• Discursive
• …In which case consider it an attempt at AWS stand-up 
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Computer Science is a “Constructed Field”
• Brought into existence from “thought stuff”
• We don’t deal with the universe in the same way as say Physics or
Biology
• No Universal Laws, no Physical Constants
• Can give us a false sense of how much control we can assert
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
The Practitioner Mindset and the evolution of Best Practice
• Clouds – And AWS in particular - are Complex
• Operating complex cloud based systems is hard
• Deal with what’s breaking or what’s going to break next
• Very focused on the here-and-now
• Not a lot of time for contemplative meditation :-/
• “Don’t just sit there, do something”…Need something actionable
• We try to deal with complexity by trying to superimpose simplicity
• Hence “Best Practices” (Or tenets, or standard operating procedures or
rules of thumb or…)
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Perspective
“Good artists borrow, great artists steal…”
• We can do worse than learn from other fields
• Surprising connections and insights can come from seemingly unrelated
places
• …which can inform and improve the state of the practitioner’s art
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Straw Poll
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Strands
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
1. The Anthropic Principle
Weak Anthropic Principle
We must be prepared to take account of the fact that our location in the
Universe is necessarily privileged to the extent of being compatible with
our existence as observers.
That we exist implies that the Universe allows for our existence.
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
The Anthropic Principle
BUT
The evidence for our existence means the Universe allows our existence,
but it doesn't mean the Universe must have unfolded exactly this way. It
doesn't mean our existence is mandatory. And it doesn't mean the
Universe must have given rise to us exactly as we are. In other words, you
cannot say “the Universe must be the way it is because we’re here.”
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
2. A War Story (Literally)
World War II. The USAF wanted to figure out where armour should be
added to fighter planes (You can’t just add armour everywhere as it makes
the fighters less manoueverable and consume more fuel and so limit
range). Data suggested that fighters that returned to base after a mission
had the most bullest holes on the rest of the plane, then the fuel system,
then the fuselage, then the engines. The air force asked how additional
armour should be distributed across where the bullet damage was the
worst.
They asked Abraham Wald, a member of the Statistical Research Group…
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Survivorship Bias
From the Wikipedia Entry for Survivorship Bias:
The logical error of concentrating on the people or things that made it past
some selection process and overlooking those that did not, typically
because of their lack of visibility.
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
3. Evolution
Struggle for existence/Survival of the fittest: The weak get weeded out, the
best survive and propagate…
It’s more nuanced than that:
• Every organism that exists today is an evolutionary winner
• But need not be tomorrow’s
• Evolution does not produce optimality in any absolute sense
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
4. The Dunning-Kruger Effect
• The scope of people’s ignorance is often invisible to them
• The least competent people can be the most confident about their
abilities
• ”Their deficits leave them with a double burden—not only does their
incomplete and misguided knowledge lead them to make mistakes but
those exact same deficits also prevent them from recognizing when
they are making mistakes and other people choosing more wisely.”
(Dunning, 2011)
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
5. Sensitive Dependence On Initial Conditions
A.k.a (more glamorously) Chaos Theory or the Butterfly Effect.
“Can a butterfly flapping its wings in Brazil cause a tornado in Texas?”
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Complex systems can be strongly nonlinear
Which is exactly what amplifies the effect of a divergence in initial
conditions…
If a best practice has been distilled from some other system’s observed
operating experience…
…How much did that system diverge from yours and how sensitive is yours
to initial conditions?
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
6. Entropy and the Second Law of Thermodynamics
Entropy: S = k ln(# of equiprobably microstates)
More colloquially: The lack of order or predictability
The Second Law of Thermodynamics (hand waving):
The entropy of a closed system never decreases.
“This law also seems plausible for software systems; as a system is
modified, its disorder, or entropy, tends to increase. This is known
as software entropy.”
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
7. Complex System Failures
A very (for me, personally) influential paper:
How Complex Systems Fail, Richard Cook
If the only thing you take away from this talk is a reference to this paper,
you’re still ahead of the game !
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Complex System Failures
18 Observations distilled from an analysis of failures in complex systems
• #3 Catastrophe requires multiple failures
• #5 Complex systems run in degraded mode
• #7 Post accident attribution to a ‘root cause’ is fundamentally wrong
• #8 Hindsight biases post-accident assessments of human performance
• #10 All practitioner actions are gambles
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Weaving The Strands
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
So remind me how this is about AWS/Best Practices/42…?
As system operators we practice our craft in a context
A context that is strongly influenced or skewed by the experience of others
We tend to extrapolate from others’ claimed experience and the prestige of
their name
And often adopt a “wisdom of crowds” best practice without a critical
analysis of why it makes sense in our specific context
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Citations
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Dunning-Kruger
Unskilled and unaware of It: How difficulties in recognizing one’s own
incompetence lead to inflated self-assessments
Kruger J, Dunning D
Journal of Personal and Social Psychology, V77(6) 1999 pp1121-1134
The Dunning-Kruger effect: On being ignorant of one’s own
ignorance
David Dunning
Advances in Experimental Social Psychology, V44 2011 pp247-296
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Dunning-Kruger
Unskilled and unaware of It: How difficulties in recognizing one’s own
incompetence lead to inflated self-assessments
Kruger J, Dunning D
Journal of Personal and Social Psychology, V77(6) 1999 pp1121-1134
The Dunning-Kruger effect: On being ignorant of one’s own
ignorance
David Dunning
Advances in Experimental Social Psychology, V44 2011 pp247-296
© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Complex System Failure
How Complex Systems Fail
Richard I. Cook, Cognitive Systems Laboratory, University of Chicago
http://web.mit.edu/2.75/resources/random/How%20Complex%20Systems
%20Fail.pdf
1 of 25

Recommended

Day 2 11.30am-12.30pm - oxford suite - yazad dalal - oracle - let's talk di... by
Day 2   11.30am-12.30pm - oxford suite - yazad dalal - oracle - let's talk di...Day 2   11.30am-12.30pm - oxford suite - yazad dalal - oracle - let's talk di...
Day 2 11.30am-12.30pm - oxford suite - yazad dalal - oracle - let's talk di...PPMA - Public Sector People Managers' Association
346 views25 slides
An Introduction to Reinforcement Learning with Amazon SageMaker by
An Introduction to Reinforcement Learning with Amazon SageMakerAn Introduction to Reinforcement Learning with Amazon SageMaker
An Introduction to Reinforcement Learning with Amazon SageMakerAmazon Web Services
533 views25 slides
An Introduction to Reinforcement Learning (December 2018) by
An Introduction to Reinforcement Learning (December 2018)An Introduction to Reinforcement Learning (December 2018)
An Introduction to Reinforcement Learning (December 2018)Julien SIMON
1.1K views25 slides
Chaos Engineering: Why Breaking Things Should Be Practiced - AWS Developer Wo... by
Chaos Engineering: Why Breaking Things Should Be Practiced - AWS Developer Wo...Chaos Engineering: Why Breaking Things Should Be Practiced - AWS Developer Wo...
Chaos Engineering: Why Breaking Things Should Be Practiced - AWS Developer Wo...Amazon Web Services
1.1K views74 slides
Closing Loops and Opening Minds: How to Take Control of Systems, Big and Smal... by
Closing Loops and Opening Minds: How to Take Control of Systems, Big and Smal...Closing Loops and Opening Minds: How to Take Control of Systems, Big and Smal...
Closing Loops and Opening Minds: How to Take Control of Systems, Big and Smal...Amazon Web Services
2.3K views51 slides
Designing for a Data-Driven Economy (AIS307) - AWS re:Invent 2018 by
Designing for a Data-Driven Economy (AIS307) - AWS re:Invent 2018Designing for a Data-Driven Economy (AIS307) - AWS re:Invent 2018
Designing for a Data-Driven Economy (AIS307) - AWS re:Invent 2018Amazon Web Services
189 views38 slides

More Related Content

Similar to An innocent's guide to ignoring best practices

Culture of Innovation - Transformation Day Philadelphia 2018 by
Culture of Innovation - Transformation Day Philadelphia 2018Culture of Innovation - Transformation Day Philadelphia 2018
Culture of Innovation - Transformation Day Philadelphia 2018Amazon Web Services
125 views36 slides
Innovation for Everyone - AWS Transformation Days Raleigh 2018.pdf by
Innovation for Everyone - AWS Transformation Days Raleigh 2018.pdfInnovation for Everyone - AWS Transformation Days Raleigh 2018.pdf
Innovation for Everyone - AWS Transformation Days Raleigh 2018.pdfAmazon Web Services
407 views39 slides
Cloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdf by
Cloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdfCloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdf
Cloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdfAmazon Web Services
581 views38 slides
Can we unwarp reality before it's too late? #IAS18 by
Can we unwarp reality before it's too late? #IAS18Can we unwarp reality before it's too late? #IAS18
Can we unwarp reality before it's too late? #IAS18Hawkeye Thompson
1.9K views78 slides
New way to learn Machine Learning with AWS DeepLens & Daniel ZivKovic by
New way to learn Machine Learning with AWS DeepLens & Daniel ZivKovicNew way to learn Machine Learning with AWS DeepLens & Daniel ZivKovic
New way to learn Machine Learning with AWS DeepLens & Daniel ZivKovicDaniel Zivkovic
237 views41 slides
Innovation at Amazon by
Innovation at AmazonInnovation at Amazon
Innovation at AmazonAmazon Web Services
2.1K views67 slides

Similar to An innocent's guide to ignoring best practices(20)

Culture of Innovation - Transformation Day Philadelphia 2018 by Amazon Web Services
Culture of Innovation - Transformation Day Philadelphia 2018Culture of Innovation - Transformation Day Philadelphia 2018
Culture of Innovation - Transformation Day Philadelphia 2018
Innovation for Everyone - AWS Transformation Days Raleigh 2018.pdf by Amazon Web Services
Innovation for Everyone - AWS Transformation Days Raleigh 2018.pdfInnovation for Everyone - AWS Transformation Days Raleigh 2018.pdf
Innovation for Everyone - AWS Transformation Days Raleigh 2018.pdf
Cloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdf by Amazon Web Services
Cloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdfCloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdf
Cloud Choices- Quantifying the Cost and Risk Implications of Cloud.pdf
Can we unwarp reality before it's too late? #IAS18 by Hawkeye Thompson
Can we unwarp reality before it's too late? #IAS18Can we unwarp reality before it's too late? #IAS18
Can we unwarp reality before it's too late? #IAS18
Hawkeye Thompson1.9K views
New way to learn Machine Learning with AWS DeepLens & Daniel ZivKovic by Daniel Zivkovic
New way to learn Machine Learning with AWS DeepLens & Daniel ZivKovicNew way to learn Machine Learning with AWS DeepLens & Daniel ZivKovic
New way to learn Machine Learning with AWS DeepLens & Daniel ZivKovic
Daniel Zivkovic237 views
Leading Your Team Through a Cloud Transformation - Virtual Transformation Day... by Amazon Web Services
Leading Your Team Through a Cloud Transformation - Virtual Transformation Day...Leading Your Team Through a Cloud Transformation - Virtual Transformation Day...
Leading Your Team Through a Cloud Transformation - Virtual Transformation Day...
Applying Principles of Chaos Engineering to Serverless (DVC305) - AWS re:Inve... by Amazon Web Services
Applying Principles of Chaos Engineering to Serverless (DVC305) - AWS re:Inve...Applying Principles of Chaos Engineering to Serverless (DVC305) - AWS re:Inve...
Applying Principles of Chaos Engineering to Serverless (DVC305) - AWS re:Inve...
Amazon Web Services6.2K views
Applying principles of chaos engineering to serverless (reinvent DVC305) by Yan Cui
Applying principles of chaos engineering to serverless (reinvent DVC305)Applying principles of chaos engineering to serverless (reinvent DVC305)
Applying principles of chaos engineering to serverless (reinvent DVC305)
Yan Cui2.8K views
Leading Your Team Through a Cloud Transformation - AWS Online Tech Talks by Amazon Web Services
Leading Your Team Through a Cloud Transformation - AWS Online Tech TalksLeading Your Team Through a Cloud Transformation - AWS Online Tech Talks
Leading Your Team Through a Cloud Transformation - AWS Online Tech Talks
Building a Culture of Innovation - AWS Partner Summit Mumbai 2018.pdf by Amazon Web Services
Building a Culture of Innovation - AWS Partner Summit Mumbai 2018.pdfBuilding a Culture of Innovation - AWS Partner Summit Mumbai 2018.pdf
Building a Culture of Innovation - AWS Partner Summit Mumbai 2018.pdf
Chaos Engineering with Kubernetes by Arun Gupta
Chaos Engineering with KubernetesChaos Engineering with Kubernetes
Chaos Engineering with Kubernetes
Arun Gupta2.1K views
Can Deep Learning Be Used to Treat Disease and Measure Product Value? (LFS310... by Amazon Web Services
Can Deep Learning Be Used to Treat Disease and Measure Product Value? (LFS310...Can Deep Learning Be Used to Treat Disease and Measure Product Value? (LFS310...
Can Deep Learning Be Used to Treat Disease and Measure Product Value? (LFS310...
Cloud choices johnenoch_theatre1_session3_1335 by John Enoch
Cloud choices johnenoch_theatre1_session3_1335Cloud choices johnenoch_theatre1_session3_1335
Cloud choices johnenoch_theatre1_session3_1335
John Enoch597 views
Enabling Sustainable Research Platforms in the Cloud by Amazon Web Services
Enabling Sustainable Research Platforms in the CloudEnabling Sustainable Research Platforms in the Cloud
Enabling Sustainable Research Platforms in the Cloud
ProTips for Scaling AWS Training to Accelerate Adoption (DVC203) - AWS re:Inv... by Amazon Web Services
ProTips for Scaling AWS Training to Accelerate Adoption (DVC203) - AWS re:Inv...ProTips for Scaling AWS Training to Accelerate Adoption (DVC203) - AWS re:Inv...
ProTips for Scaling AWS Training to Accelerate Adoption (DVC203) - AWS re:Inv...
Amazon Web Services6.1K views
AI Powered Conversational Interfaces for Personalized Learning & Chatbots by Amazon Web Services
AI Powered Conversational Interfaces for Personalized Learning & ChatbotsAI Powered Conversational Interfaces for Personalized Learning & Chatbots
AI Powered Conversational Interfaces for Personalized Learning & Chatbots
What IT Transformation Really Means for the Enterprise by Tom Laszewski
What IT Transformation Really Means for the EnterpriseWhat IT Transformation Really Means for the Enterprise
What IT Transformation Really Means for the Enterprise
Tom Laszewski146 views

More from AWS User Group Bengaluru

Demystifying identity on AWS by
Demystifying identity on AWSDemystifying identity on AWS
Demystifying identity on AWSAWS User Group Bengaluru
440 views108 slides
AWS Secrets for Best Practices by
AWS Secrets for Best PracticesAWS Secrets for Best Practices
AWS Secrets for Best PracticesAWS User Group Bengaluru
931 views21 slides
Cloud Security by
Cloud SecurityCloud Security
Cloud SecurityAWS User Group Bengaluru
1.9K views133 slides
Lessons learnt building a Distributed Linked List on S3 by
Lessons learnt building a Distributed Linked List on S3Lessons learnt building a Distributed Linked List on S3
Lessons learnt building a Distributed Linked List on S3AWS User Group Bengaluru
312 views27 slides
Medlife journey with AWS by
Medlife journey with AWSMedlife journey with AWS
Medlife journey with AWSAWS User Group Bengaluru
336 views17 slides
Building Efficient, Scalable and Resilient Front-end logging service with AWS by
Building Efficient, Scalable and Resilient Front-end logging service with AWSBuilding Efficient, Scalable and Resilient Front-end logging service with AWS
Building Efficient, Scalable and Resilient Front-end logging service with AWSAWS User Group Bengaluru
443 views33 slides

More from AWS User Group Bengaluru(20)

Building Efficient, Scalable and Resilient Front-end logging service with AWS by AWS User Group Bengaluru
Building Efficient, Scalable and Resilient Front-end logging service with AWSBuilding Efficient, Scalable and Resilient Front-end logging service with AWS
Building Efficient, Scalable and Resilient Front-end logging service with AWS
Building Efficient, Scalable and Resilient Front-end logging service with AWS by AWS User Group Bengaluru
Building Efficient, Scalable and Resilient Front-end logging service with AWSBuilding Efficient, Scalable and Resilient Front-end logging service with AWS
Building Efficient, Scalable and Resilient Front-end logging service with AWS

Recently uploaded

Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha... by
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...ShapeBlue
138 views18 slides
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading... by
Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading...The Digital Insurer
86 views52 slides
Ransomware is Knocking your Door_Final.pdf by
Ransomware is Knocking your Door_Final.pdfRansomware is Knocking your Door_Final.pdf
Ransomware is Knocking your Door_Final.pdfSecurity Bootcamp
90 views46 slides
Kyo - Functional Scala 2023.pdf by
Kyo - Functional Scala 2023.pdfKyo - Functional Scala 2023.pdf
Kyo - Functional Scala 2023.pdfFlavio W. Brasil
449 views92 slides
CloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlue by
CloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlueCloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlue
CloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlueShapeBlue
94 views13 slides
Extending KVM Host HA for Non-NFS Storage - Alex Ivanov - StorPool by
Extending KVM Host HA for Non-NFS Storage -  Alex Ivanov - StorPoolExtending KVM Host HA for Non-NFS Storage -  Alex Ivanov - StorPool
Extending KVM Host HA for Non-NFS Storage - Alex Ivanov - StorPoolShapeBlue
84 views10 slides

Recently uploaded(20)

Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha... by ShapeBlue
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...
ShapeBlue138 views
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading... by The Digital Insurer
Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading...
CloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlue by ShapeBlue
CloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlueCloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlue
CloudStack Managed User Data and Demo - Harikrishna Patnala - ShapeBlue
ShapeBlue94 views
Extending KVM Host HA for Non-NFS Storage - Alex Ivanov - StorPool by ShapeBlue
Extending KVM Host HA for Non-NFS Storage -  Alex Ivanov - StorPoolExtending KVM Host HA for Non-NFS Storage -  Alex Ivanov - StorPool
Extending KVM Host HA for Non-NFS Storage - Alex Ivanov - StorPool
ShapeBlue84 views
Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava... by ShapeBlue
Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava...Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava...
Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava...
ShapeBlue101 views
Initiating and Advancing Your Strategic GIS Governance Strategy by Safe Software
Initiating and Advancing Your Strategic GIS Governance StrategyInitiating and Advancing Your Strategic GIS Governance Strategy
Initiating and Advancing Your Strategic GIS Governance Strategy
Safe Software140 views
Setting Up Your First CloudStack Environment with Beginners Challenges - MD R... by ShapeBlue
Setting Up Your First CloudStack Environment with Beginners Challenges - MD R...Setting Up Your First CloudStack Environment with Beginners Challenges - MD R...
Setting Up Your First CloudStack Environment with Beginners Challenges - MD R...
ShapeBlue132 views
Backup and Disaster Recovery with CloudStack and StorPool - Workshop - Venko ... by ShapeBlue
Backup and Disaster Recovery with CloudStack and StorPool - Workshop - Venko ...Backup and Disaster Recovery with CloudStack and StorPool - Workshop - Venko ...
Backup and Disaster Recovery with CloudStack and StorPool - Workshop - Venko ...
ShapeBlue144 views
CloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&T by ShapeBlue
CloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&TCloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&T
CloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&T
ShapeBlue112 views
Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And... by ShapeBlue
Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And...Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And...
Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And...
ShapeBlue63 views
Why and How CloudStack at weSystems - Stephan Bienek - weSystems by ShapeBlue
Why and How CloudStack at weSystems - Stephan Bienek - weSystemsWhy and How CloudStack at weSystems - Stephan Bienek - weSystems
Why and How CloudStack at weSystems - Stephan Bienek - weSystems
ShapeBlue197 views
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue by ShapeBlue
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlueWhat’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue
ShapeBlue222 views
VNF Integration and Support in CloudStack - Wei Zhou - ShapeBlue by ShapeBlue
VNF Integration and Support in CloudStack - Wei Zhou - ShapeBlueVNF Integration and Support in CloudStack - Wei Zhou - ShapeBlue
VNF Integration and Support in CloudStack - Wei Zhou - ShapeBlue
ShapeBlue163 views
iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas... by Bernd Ruecker
iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas...iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas...
iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas...
Bernd Ruecker50 views
Updates on the LINSTOR Driver for CloudStack - Rene Peinthor - LINBIT by ShapeBlue
Updates on the LINSTOR Driver for CloudStack - Rene Peinthor - LINBITUpdates on the LINSTOR Driver for CloudStack - Rene Peinthor - LINBIT
Updates on the LINSTOR Driver for CloudStack - Rene Peinthor - LINBIT
ShapeBlue166 views
Confidence in CloudStack - Aron Wagner, Nathan Gleason - Americ by ShapeBlue
Confidence in CloudStack - Aron Wagner, Nathan Gleason - AmericConfidence in CloudStack - Aron Wagner, Nathan Gleason - Americ
Confidence in CloudStack - Aron Wagner, Nathan Gleason - Americ
ShapeBlue88 views

An innocent's guide to ignoring best practices

  • 1. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. BENGALURU
  • 2. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. An Innocent’s Guide To Ignoring Best Practices... AMIT GARDE, VP Engineering, Swiggy | 6 October 2018
  • 3. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Preliminaries • What this talk is *not* going to be… • Prescriptive • A laundry list of things to do or to not do • What this talk is *trying* to be • A perspective on how “Best Practice” emerges, why a grain of salt is useful, and why we could usefully borrow ideas from other fields of human endeavour • What this talk *might* end up being • Whimsical • Discursive • …In which case consider it an attempt at AWS stand-up 
  • 4. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Computer Science is a “Constructed Field” • Brought into existence from “thought stuff” • We don’t deal with the universe in the same way as say Physics or Biology • No Universal Laws, no Physical Constants • Can give us a false sense of how much control we can assert
  • 5. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. The Practitioner Mindset and the evolution of Best Practice • Clouds – And AWS in particular - are Complex • Operating complex cloud based systems is hard • Deal with what’s breaking or what’s going to break next • Very focused on the here-and-now • Not a lot of time for contemplative meditation :-/ • “Don’t just sit there, do something”…Need something actionable • We try to deal with complexity by trying to superimpose simplicity • Hence “Best Practices” (Or tenets, or standard operating procedures or rules of thumb or…)
  • 6. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Perspective “Good artists borrow, great artists steal…” • We can do worse than learn from other fields • Surprising connections and insights can come from seemingly unrelated places • …which can inform and improve the state of the practitioner’s art
  • 7. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Straw Poll
  • 8. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Strands
  • 9. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 1. The Anthropic Principle Weak Anthropic Principle We must be prepared to take account of the fact that our location in the Universe is necessarily privileged to the extent of being compatible with our existence as observers. That we exist implies that the Universe allows for our existence.
  • 10. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. The Anthropic Principle BUT The evidence for our existence means the Universe allows our existence, but it doesn't mean the Universe must have unfolded exactly this way. It doesn't mean our existence is mandatory. And it doesn't mean the Universe must have given rise to us exactly as we are. In other words, you cannot say “the Universe must be the way it is because we’re here.”
  • 11. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 2. A War Story (Literally) World War II. The USAF wanted to figure out where armour should be added to fighter planes (You can’t just add armour everywhere as it makes the fighters less manoueverable and consume more fuel and so limit range). Data suggested that fighters that returned to base after a mission had the most bullest holes on the rest of the plane, then the fuel system, then the fuselage, then the engines. The air force asked how additional armour should be distributed across where the bullet damage was the worst. They asked Abraham Wald, a member of the Statistical Research Group…
  • 12. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Survivorship Bias From the Wikipedia Entry for Survivorship Bias: The logical error of concentrating on the people or things that made it past some selection process and overlooking those that did not, typically because of their lack of visibility.
  • 13. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 3. Evolution Struggle for existence/Survival of the fittest: The weak get weeded out, the best survive and propagate… It’s more nuanced than that: • Every organism that exists today is an evolutionary winner • But need not be tomorrow’s • Evolution does not produce optimality in any absolute sense
  • 14. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 4. The Dunning-Kruger Effect • The scope of people’s ignorance is often invisible to them • The least competent people can be the most confident about their abilities • ”Their deficits leave them with a double burden—not only does their incomplete and misguided knowledge lead them to make mistakes but those exact same deficits also prevent them from recognizing when they are making mistakes and other people choosing more wisely.” (Dunning, 2011)
  • 15. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 5. Sensitive Dependence On Initial Conditions A.k.a (more glamorously) Chaos Theory or the Butterfly Effect. “Can a butterfly flapping its wings in Brazil cause a tornado in Texas?”
  • 16. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Complex systems can be strongly nonlinear Which is exactly what amplifies the effect of a divergence in initial conditions… If a best practice has been distilled from some other system’s observed operating experience… …How much did that system diverge from yours and how sensitive is yours to initial conditions?
  • 17. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 6. Entropy and the Second Law of Thermodynamics Entropy: S = k ln(# of equiprobably microstates) More colloquially: The lack of order or predictability The Second Law of Thermodynamics (hand waving): The entropy of a closed system never decreases. “This law also seems plausible for software systems; as a system is modified, its disorder, or entropy, tends to increase. This is known as software entropy.”
  • 18. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. 7. Complex System Failures A very (for me, personally) influential paper: How Complex Systems Fail, Richard Cook If the only thing you take away from this talk is a reference to this paper, you’re still ahead of the game !
  • 19. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Complex System Failures 18 Observations distilled from an analysis of failures in complex systems • #3 Catastrophe requires multiple failures • #5 Complex systems run in degraded mode • #7 Post accident attribution to a ‘root cause’ is fundamentally wrong • #8 Hindsight biases post-accident assessments of human performance • #10 All practitioner actions are gambles
  • 20. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Weaving The Strands
  • 21. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. So remind me how this is about AWS/Best Practices/42…? As system operators we practice our craft in a context A context that is strongly influenced or skewed by the experience of others We tend to extrapolate from others’ claimed experience and the prestige of their name And often adopt a “wisdom of crowds” best practice without a critical analysis of why it makes sense in our specific context
  • 22. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Citations
  • 23. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Dunning-Kruger Unskilled and unaware of It: How difficulties in recognizing one’s own incompetence lead to inflated self-assessments Kruger J, Dunning D Journal of Personal and Social Psychology, V77(6) 1999 pp1121-1134 The Dunning-Kruger effect: On being ignorant of one’s own ignorance David Dunning Advances in Experimental Social Psychology, V44 2011 pp247-296
  • 24. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Dunning-Kruger Unskilled and unaware of It: How difficulties in recognizing one’s own incompetence lead to inflated self-assessments Kruger J, Dunning D Journal of Personal and Social Psychology, V77(6) 1999 pp1121-1134 The Dunning-Kruger effect: On being ignorant of one’s own ignorance David Dunning Advances in Experimental Social Psychology, V44 2011 pp247-296
  • 25. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Complex System Failure How Complex Systems Fail Richard I. Cook, Cognitive Systems Laboratory, University of Chicago http://web.mit.edu/2.75/resources/random/How%20Complex%20Systems %20Fail.pdf

Editor's Notes

  1. Tie in to “Ignoring Best Practices”…practitioner actions need to be grounded in principles and understanding, not blind obedience to a set of rules. I’m deliberately not going to get into the details of a specific supposedly best practice.
  2. In *principle* we may not think we have the control right now but we do believe that is is possible in principle, if only we did better in some way
  3. The intent in the next part will be to bring in some strands,
  4. How many here from a non CS/Engg background ? [Ramble a bit] How many have any familiarity with the hard sciences Physics, Biology etc post high school? How many with any social science ? [If none, or negligible as expected, this proves the point raffle. Else pleasantly surprised blah blah
  5. Has
  6. So if you’re running an AWS based infrastructure and it’s up, it doesn’t mean there was anything particularly great about what you did
  7. Wald recommended adding armour to the engines The one's you're missing are the ones who didn't make it back
  8. Why is this relevant? If your system has emerged after many iterations and accreted complexity there is a belief it must be better, it’s been running for so long. Not necessarily true. Also if the environment changes or there is the tech equivalent of a ‘mutation’ (spectre/meltdown and the firmware patches for them)
  9. Dilbert Fans? Pointy Headed Boss reference
  10. If you’re borrowing best practices from other people, the sensitivity to initial condition might mean that extrapolating from their experience is meaningless. Best Practice derives from (someone’s) experience, and if that someone’s starting state is sufficiently different from yours even if slightly, the sensitive dependence can mean applying their experience in your context buys you nothing – or worse
  11. If you’re borrowing best practices from other people, the sensitivity to initial condition might mean that extrapolating from their experience is meaningless. Best Practice derives from (someone’s) experience, and if that someone’s starting state is sufficiently different from yours even if slightly, the sensitive dependence can mean applying their experience in your context buys you nothing – or worse
  12. 10 configuration parameters; even if each is a simple on/off binary, that's already 1024...
  13. If the only thing you take away from this talk is the reference for this paper, that will have been worth it.
  14. A theme should have emerged by now...
  15. If the only thing you take away from this talk is the reference for this paper, that will have been worth it.