Key Trends Shaping Cloud Infrastructure and Edge Infrastructure

A talk at Edgewise 2024 about key trends and also a personal account of my journey.

https://www.fullstaq.com/edgecase-2024

1. TRENDS SHAPING THE
FUTURE OF CLOUD
NATIVE AND EDGE
CHERYL HUNG, OICHERYL.COM
EDGECASE, AMERSFOORT, 24 SEPTEMBER2024

2. CHERYL HUNG
Sr Director
Founder, Cloud
Native London

3. Vendor-owned Open Source
The Age of AI
Building the Future, Together

4. BUILDING THE
FUTURE,
TOGETHER

6. FINDING
TECHNOLOGY

7. VENDOR-OWNED
OPEN SOURCE

8. OPEN SOURCE FLYWHEEL

9. OPEN SOURCE FLYWHEEL
•Speed
Innovation
•Drive
Adoption
•Improve
Reliability

10. OPEN SOURCE FLYWHEEL
•Speed
Innovation
•Drive
Adoption
•Improve
Reliability
•Pay for support
•Open Core
•SaaS

11. OPEN SOURCE RELICENSING
Oct 2018, MongoDB AGPLv3 > SSPL
Feb 2019, Redis Modules under Redis Source Available License
Jan 2021, Elasticsearch Apache2 > SSPL, Elastic License
Apr 2021, Grafana Apache2 > AGPLv3
Aug 2023, Terraform MPL 2.0 > Business Source License

12. linkerd.io/2024/02/21/announcing-linkerd-2.15/
“While Linkerd will always be open
source, as of Linkerd 2.15,
producing stable Linkerd releases
will be in the hands of the vendor
community”

13. MANAGE YOUR OPEN SOURCE EXPOSURE

14. MANAGE YOUR OPEN SOURCE EXPOSURE
Which open-source license does it use?

15. MANAGE YOUR OPEN SOURCE EXPOSURE
Which open-source license does it use?
Can you automate compliance checks?

16. MANAGE YOUR OPEN SOURCE EXPOSURE
Which open-source license does it use?
Can you automate compliance checks?
Who’s behind the open source?

17. MANAGE YOUR OPEN SOURCE EXPOSURE
Which open-source license does it use?
Can you automate compliance checks?
Who’s behind the open source?
What’s the governance policy?

18. FINDING
COMMUNITY

19. THE AGE OF AI

20. ❑Yes, in production
❑Yes, but not in production
❑No, but considering in the
next 12 months
❑No, not considering
DOES YOUR ORG RUN AI WORKLOADS?

21. >50% of est Data Center capacity
dedicated for AI Training and Inferencing
Data Proliferation - Global Data Use to
Reach 175 Zetta Bytes by 2025
Sustainability - Infrastructure Consumes
3% of Global Energy

22. Training
(GPUs, TPUs)
Inference (CPUs)
*Source: Schneider Electric – AI Disruption Report 2023
15%
85%

23. “The era of large models is
over, and the focus will now
turn to specializing and
customizing these models.
Real value is unlocked only
when these models are tuned
on customer and domain
specific data.”
Sam Altman, CEO OpenAI

24. Lab Reference
Range ID
70B
Financial
Sentiment
Analysis 13B
Code
completion and
creation
7-70B

25. NEED FOR SPECIALIZED COMPUTE AT LOW
TCO (PERF/WATT) AT SCALE
Moore’s Law Can’t Keep Up
Sustainability - Infrastructure Consumes
3% of Global Energy

26. Internet connected IoT devices expected to
double to 30 billion by 2030

28. WHAT IS CLOUD NATIVE AI?

29. WHAT IS CLOUD NATIVE AI?
Leveraging cloud computing (CPUs, GPUs),
network and storage, with isolation and controlled
sharing mechanisms, to deploy, run, and scale AI
workloads.

30. WHAT IS CLOUD NATIVE AI?
Leveraging cloud computing (CPUs, GPUs),
network and storage, with isolation and controlled
sharing mechanisms, to deploy, run, and scale AI
workloads.
WHY?

31. WHAT IS CLOUD NATIVE AI?
Leveraging cloud computing (CPUs, GPUs),
network and storage, with isolation and controlled
sharing mechanisms, to deploy, run, and scale AI
workloads.
WHY?
Accelerate performance, reduce cost

33. •Resource Provisioning
•Cost control
•Observability
•Disaster Recovery
•Security
•Sustainability
CHALLENGES

35. SOLUTIONS
LANDSCAPE

36. ECOSYSTEM
•AI Engineering
•Safety by Design
•Education for kids and students

37. FINDING
THE FUTURE

38. THANK YOU
OICHERYL.COM

39. RESOURCES
•CNCF Annual Survey 2023: cncf.io/reports/cncf-annual-
survey-2023/
•Cloud Native AI White Paper: cncf.io/reports/cloud-native-
artificial-intelligence-whitepaper/
•Health of Linkerd project: github.com/cncf/toc/issues/1262