This presentation explains what serverless is all about, explaining the context from Devs & Ops points of view, and presenting the various ways to achieve serverless (Functions a as Service, BaaS....). It also presents the various competitors on the market and demo one of them, openfaas. Finally, it enlarges the pictures, positionning serverless, combined with Edge computing & IoT, as a valuable triptic cloud vendors are leveraging on top of, to create end-to-end offers.

1. @adrienblind
Introduction to Serverless
Adrien Blind

2. @adrienblind
Are you a server hugger?

3. @adrienblind
What do you really want
Deliver rapidely and flowly valuable apps
for the business

4. @adrienblind
What do you really want
Cloud
Agile
DevOps
Microservice
architecture
Craftsmanship
Deliver rapidely and flowly valuable apps
for the business

5. @adrienblind
What do you really want
On demand
Pay as you go
ElasticCloud
Deliver rapidely and flowly valuable apps
for the business
Agile
DevOps
Microservice
architecture
Craftsmanship

6. @adrienblind
A new kid on the block ?

7. @adrienblind
A new “cloudy” kid on the block

8. @adrienblind
A single definition for Serverless?
“Serverless computing is a cloud computing execution model in which the cloud provider dynamically manages the
allocation of machine resources”
- Wikipedia
“Serverless computing refers to the concept of building and running applications that do not require server
management. It describes a finer-grained deployment model where applications, bundled as one or more
functions, are uploaded to a platform and then executed, scaled, and billed in response to the exact demand
needed at the moment.”
- CNCF foundation whitepaper on Serverless
“Serverless architectures refer to applications that significantly depend on third-party services (knows as Backend
as a Service or "BaaS") or on custom code that's run in ephemeral containers (Function as a Service or "FaaS") [...].
such architectures remove the need for the traditional 'always on' server system sitting behind an application.”
- Mike Roberts, martinfowler.com (2016)
“If your PaaS can efficiently start instances in 20ms that run for half a second, then call it serverless.”
- Adrian Cockroft (2016)

9. @adrienblind
What are we talking about
❏ Dev/App perspective
Arch. design & granularity? Ephemeral apps? PaaS pattern?
Mostly all of them!
❏ Ops/platform perspective: infrastructure abstraction
Elastic → autoscaling
On-demand → boot in 20ms
Pay-as-you-go → Scale to zero
Cloud managed services
& … NoOps?

10. @adrienblind
2 major serverless tendencies
❏ BaaS, aka Backend-as-a-Service
&
❏ FaaS, aka Functions-as-a-Service

11. @adrienblind
❏ Checkout first what we do already have today…
we’ll dig into BaaS & FaaS in few minutes ;)

12. @adrienblind
Cloud: IaaS
❏ End-product is almost the same (virtual machines)
❏ Facilitates lift & shift
❏ Interaction with the product change: cloud principles, automation/infra-as-
code
❏ Stimulated the commoditization trend: ops no longer deliver per-app
infrastructure architectures
❏ Starts to shift resiliency & scalability topics from infra to app
Ok… But not Serverless at all.

13. @adrienblind
❏ Containers isolate apps from their subsequent OS
❏ Orchestrators enables to consider a pool of OS as a global capacity (CaaS)
❏ Possible Auto-scalability but mostly based on infrastructure metrics (CPU)
❏ However, pricing model mostly related to subsequent infrastructure used (cluster nodes)
❏ Containers commonly associated to finer app granularity/microservices, which may lead to more
efficient startup time
You now start observing few Serverless characteristics
Cloud: CaaS
+
Create more
abstraction from IaaS

14. @adrienblind
Cloud: PaaS or PaaS?
❏ PaaS stands for Platform-as-a-Service
❏ Application PaaS (aPaaS) pioneer: Google App engine
❏ xPaaS = managed services (DBaaS, object storage, etc.)
❏ Complete abstraction of the service from infrastructure
❏ Pricing model not related to infrastructure
❏ Autoscaling & resilient by design
Quite Serverless compliant, nope ?

15. @adrienblind
❏ BaaS, aka Backend-as-a-Service

16. @adrienblind
Backend-as-a-Service architecture
BaaS → no!
PaaS
CaaS
IaaS

17. @adrienblind
❏Not a cloud service offer, it’s an architecture concept
❏Trend initially launched for mobile apps
❏Build apps directly leveraging on PaaS services
❏Security paradigms shifts
From perimetric to embedded
Auth/Authz/Cipher everything (your backend is exposed, mate…)
Backend-as-a-Service architecture
Authentication
Object Storage
Databases
(ex. DynamoDB, etc.)
Mobile app
Or Browser

18. @adrienblind
❏ FaaS, aka Functions-as-a-Service

19. @adrienblind
Client
Gateway
FaaS
platform
FaaS compute capacity abstracted from app perspective
Cloud: FaaS
❏ Ephemeral: platform waits requests...

20. @adrienblind
Client
Instanciated
function
(your code)
Gateway
FaaS
platform
FaaS compute capacity abstracted from app perspective
Cloud: FaaS
❏ Ephemeral: platform waits requests and
triggers function on demand, which “lives”
the time to deliver the result: your code is
not always-on, waiting calls!

21. @adrienblind
❏ Ephemeral: platform waits requests and
triggers function on demand, which “lives”
the time to deliver the result: your code is
not always-on, waiting calls!
❏ Dynamic scalability & resilience provided
by the platform: more calls, more
instances
❏ Extremely fine grained pay-per-use on
public cloud: per-call costs
Client
Instanciated
function
(your code)
Instanciated
function
(your code)
Instanciated
function
(your code)
Gateway
FaaS
platform
FaaS compute capacity abstracted from app perspective
Cloud: FaaS

22. @adrienblind
❏Beware of design constraints applying to your app
❏ Service/function granularity
❏ Stateless services, use distributed cache, etc.
❏ Small wake up time
❏ No long-running services
❏Like for the Virtual machines & container history, shortly after their venue
came the orchestration challenge… Deal with function graph calls (ex.
Amazon Step)
❏Possible Vendor lock-in: check out serverless.io to get an abstraction fmk ?
❏Testing → you must deploy on the platform everytime
❏Adapt DevOps practices: monitoring, deployment, versioning...
FaaS constraints

23. @adrienblind
All major cloud vendors have products
Alternatively you can deploy your own FaaS fmk
You may leverage on existing CaaS and put value on top of it
• Container encapsulation of the function
• Kubernetes orchestration
FaaS on public cloud vs FaaS on premises
Amazon Google Azure OVH
OpenWhisk OpenFaaS OracleFN

24. @adrienblind
❏ OpenFaas: an example FaaS platform

25. @adrienblind
Introduction to OpenFaaS
 Originally a cool hack built by docker captain Alex Ellis for
Dockercon conference. A new, but real product now.
 Enable to trigger functions enclosed in docker images. So:
You can put any app/language, till there is a binary to start
Works Linux and Windows OS
On premises, on cloud, hybridated…
 https://www.openfaas.com

26. @adrienblind
OpenFaas leverage on containers & orchestration
Illustration from openfaas.com

27. @adrienblind
An example platform: introduction to OpenFaaS
 API gateway
Route inbound requests to your functions
Collect metrics through Prometheus
Autoscales modifying service replicas counts
Offers a convenient UI + endpoints for the CLI

28. @adrienblind
An example platform: introduction to OpenFaaS
 Function watchdog
A tiny HTTP server, enclosed along with your app in the docker image
Receive request from the API Gateway, triggers your app
Provide args and catch result through STDIN/STDOUT
Now, you know why you can enclose almost everything ;)
Illustration from openfaas.comIllustration from openfaas.com

29. @adrienblind
$ faas-cli new helloworld --lang python
The only line of code you’ll write in the function template
print("Hello, World!")
$ faas-cli build -f helloworld.yml
$ faas-cli push -f helloworld.yml
$ faas-cli deploy -f helloworld.yml
$ curl http://faas.me/function/helloworld
Hello, World!
“Hello, World!” example With OpenFaaS

30. @adrienblind
Benefits
 Leverage on your CaaS platform for serverless &
traditional usecase at the same time
 Leverage on top of Docker paradigms
Universality of languages/app supported
Easy CD/CI, automation
 Do not reinvent the wheel 

31. @adrienblind
❏ Sum Up

32. @adrienblind
❏ Serverless is an architecture
trend, not just a new cloud
service offer (aka FaaS)
❏ A serverless app is a mashup of
value-added, managed services:
aPaaS, xPaaS, FaaS…
❏ Fits today’s architecture stakes:
cloud native apps, 12 factors...
Serverless key aspects from Dev/App perspective
Schema from martinfowler.com
Devs gain a greater productivity, refocusing on business valuable code

33. @adrienblind
Ops gets more efficiency and cost-saving,
offloading several concerns to platform
❏ You no longer manage infrastructure aspects directly: Auto-
scalability & resilience provided by design
❏ Close but not fully NoOps (monitor app!)
❏ Cloud promise at its finest “resolution”
On-demand, Elastic, Pay-Per-Use
Serverless key aspects from Ops/platform perspective

34. @adrienblind
Focus on value & better TTM
but support platform constraints
More flexibility, more tuning
capacity… but assume plumbing
Functions
Serverless key aspects
ABSTRACT
FOCUSINFRA
Microservices
Monoliths
CaaS
IaaS
FaaS
& PaaS

35. @adrienblind
❏ Small microservices
❏ Stream processing
❏ Event-driven-programming
❏ Batch / Scheduled tasks
❏ May even replace some old compute grids ;)
❏ …
Some usecases

36. @adrienblind
Finally, THE silver bullet?
Highly valuable, but does not fit every usecases !

37. @adrienblind
❏ Extra: enlarge the picture

38. @adrienblind
IoT & Serverless
❏ IoT generates large loads of small & basic-to-process
events, in huge quantity
❏ It calls for an event-driven programming approach
❏ … which fits well with the idea of simple, elementary
functions of Serverless/FaaS computing
Serverless
+
IoT

39. @adrienblind
Edge Computing & Serverless
❏ Google Trends graphs for “Serverless” & “Edge computing” terms
❏ Beware, scales are not the same however ;)
❏ Anyway, an interesting correlation to notice, isn’t it ?

40. @adrienblind
WTF with Edge computing?
❏ Offload computing tasks close to the
data, at the boarder of the network / out
from the cloud
❏ Example, precompute face recognition
close to a camera, to avoid uploading the
whole video flow to the cloud
❏ Particularly valuable in an IoT landscape
CLOUD

41. @adrienblind
Major cloud vendors are building their strategy on top of the following triptic,
to unleash their service from the cloud and provide an end-to-end offer
For instance: Azure IoT Edge / Sphere, AWS Greengrass...
Unleash from the cloud
Serverless
Architecture
Edge
Computing
Internet Of
Things

42. @adrienblind
There are 2 kind of spurs my friend.
Those that do serverless, those that don’t.