Il existe de nombreuses possibilités pour exécuter une application ou du code dans Azure. Nous examinerons les différentes options afin de les positionner les unes par rapport aux autres : Machines virtuelles, conteneurs, services, serverless.
Tour de France Azure PaaS 2/7 Exécuter une application
1. Tour de France
Azure PaaS
Lyon 3 Avril, Issy-les-Moulineaux 5 Avril, Lille 10 Avril,
Bordeaux 12 Avril, Toulouse 24 Avril
Aix-en-Provence 26 Avril, Nantes 3 Mai
21. Expanding Developer Experiences
New workload APIs for easy integration
Ease to take advantage of low-priority batch
D Nc …A
Azure Batch
Infra
Azure Batch
Batch AI
Training R Studio
Batch
Rendering
Platform
Workload
Interfaces
More…
Jobs
With Hybrid Use Benefits
Low-Priority Batch
Hybrid licensing for Windows with SA
Windows discount increases to 80%
Preemptible VMs with up to 80% fixed
All Batch VM sizes and regions
Mix low-pri and dedicated priority VMs in a pool
23. Azure Container Instances (ACI)
Easily run containers on Azure with a single command
PREVIEW
Cloud-scale
container capacity
Hyper-visor
isolation
Start using
containers right away
27. Azure Container Service (AKS)
Simplify the deployment, management, and
operations of Kubernetes
Work how you
want with open-
source APIs
Scale and run
applications with
confidence
Focus on your
containers not the
infrastructure
61. Fault-tolerant E-Commerce with personalized recommendations
Traffic Manager
Power BI
Recommendations API
East West
Browser
Application
Insights
Azure Active
Directory B2C
Priority Traffic-Routing Method
CDN
Redis
Cache
SQL Database
Product Catalog
Orders
Product Images
Site Content
Search
Commerce
Site
Queue
Credit Card
Transactions
WebJob
Credit Card
Processing
Third party
Payment
Commerce
Site
Queue
Credit Card
Transactions
WebJob
Credit Card
Processing
Third party
Payment
Redis
Cache
SQL Database
Product Catalog
Orders
Product Images
Site Content
SearchPower BI
62. Real-time Inventory integration supporting IoT
API
Management
Browser
Commerce Site Queue
Credit Card
Transactions
WebJob
Credit Card
Processing
Third party
Payment
Azure Active
Directory B2C
Application
Insights
CDN
Redis
Cache
SQL Database
Product Catalog
Orders
Product
Images
Site Content
Search
IoT Hub
Stream
Analytics
Service Bus
Service Fabric
Blob Storage
Archive
DocumentDB
Beacon
Sensor
Asset
Edge Gateway
Map Key
Retail Store
Location
86. Logic Apps is
Powerful Integration
Faster integration using innovative Visual Designer
Easy workflow creation with triggers and actions
Mashup applications, data and services
Built for mission critical Enterprise Integration
Create, deploy, manage and monitor
Really the basic building block of all other compute services
Valuable for cases where you’re simply lifting-and-shifting an existing workload to the cloud and want to maintain the existing topology.
Huge variety of OS images available, including all the popular Linux distros
Support for many popular configuration tools, including Chef and Puppet.
Available in many different sizes to fit your workload, including the recently announced N-series which offer Nvidia GPUs for graphically intense workloads.
And of course, since VMs make up the basic building blocks in our Compute story, many people have built interesting things out of them, which are available in the marketplace…
Tesla P40 for AI & Deep Learning training and inference Workloads single precision performance (FP32) and larger models.
Tesla P100 for traditional HPC workloads for over 2 – 4x compared to current generation workloads, double precision. (FP64)
Significant cost savings
Save up to 72% on pay-as-you-go (PAYG) prices—with one-year or three-year terms on Windows and Linux VMs. When you combine the cost savings inherent to Azure RIs with the added value of the Azure Hybrid Benefit, you can save up to 82%.
Lower your TCO by combining RIs with PAYG to manage costs across predictable and variable workloads.
Budget predictability and prioritized compute capacity
Improve budgeting and forecasting with a single up-front payment.
Select your preferred VM series and Azure region to optimize capacity for your workloads.
Flexibility to modify reservations and simplicity of purchase
Select and purchase RIs in three easy steps. Specify your region, virtual machine series, and term (one or three years) and leave the rest to us. Plan, implement, and execute against your desired workloads with prioritized datacenter capacity around the globe—ensure resources are available when and where you need them.
While RIs require making upfront commitments on compute capacity, they also provide flexibility should your business needs change. Easily exchange or cancel RIs with the buyback program.
Azure Reserved VM Instances enable customers to reserve compute capacity to prioritize workloads when and where they need it most. With a 72% cost savings over on-demand pricing, reserved instances offer price predictability because they are purchased in 1- and 3-year terms with a single up-front payment.
What’s more, Windows Server customers can save up to 82% with AHUB. Azure Reserved VM Instances offer the most affordable and flexible RIs with prioritized compute capacity on the market.
So, VMs are great… but managing a lot of them individually can be hard. That’s where VM scale sets comes in..
The goal here is to enable management of like VMs as a group and make it easy to scale them in/out based on demand.
There are many different ways to deploy software to VM scale sets. I won’t go into any great detail here, but just to give you a flavor…
If you’re simply looking to deploy an existing application off-the-shelf, you can use the Azure marketplace
Alternatively, you can use what we call a VM extension, which is a task triggered inside the VM after the base image has booted, where you can run arbitrary scripts.
You can also deploy a custom VM image that you’ve created with a tool like Packer
So with that little bit of context on VM scale sets, let’s start slowly moving up the stack…
While you can use VMSS directly, we expect most customers will interact with them through a higher-level abstraction.
One of those is Azure Batch, which enables highly-parallelizable job processing. You provide it with your job, configure the amount of compute you want to throw at it, and Batch will both schedule it efficiently and restart failed tasks as required.
Common scenarios include financial risk modeling, climate analysis, and media encoding
*App Service on Azure Stack is in Preview as of July 2017.
Momentum data captured June 2017.
S4, Solution Specialist Sales Summit
Abstraction of servers, infrastructure and configuration of operating system
Event-driven scale
Sub-second billing
Stateless
Serverless compute is a fully managed service. Some refer to it as Functions as a Service
OS and Framework patching is performed for you
There is zero administrative tasks and no need to manage any infrastructure
You just deploy your code (function) and it runs
Your code runs within seconds and for very short period of time
Serverless compute scales quickly (almost instantly) and vastly
Automatically scales within seconds
No scale configuration is required (there is no way to configure scale or limits)
Scales to match any given workload. Scales from zero to handle tens of thousands concurrent functions invocations within seconds
Pay only for the time your code is running
Serverless compute reacts to events
React, in near real-time, to events and triggers
Triggered by virtually any event from Azure service or 3rd party services
Mono-lithic application -> loosely coupled components
Events happening async outside client application
Build/use micro-functionality to react to a single event
“Pinnacle of PaaS compute”
Not just hardware “servers”, but software servers are also managed for you
Focus on business logic, not solving technical problems not core to business
Lower effort to get started makes it easier to experiment (bots, etc.)
Chris
This is the whole (top level) Dev Ops story where Dev build wounderful apps, using CI they mange their code, builds, and releases. Connecting to a repository, they can deploy to Azure Websites, build in the cloud, deploy to a slot to test and validate, swap into production, deployed across the globe as needed, monitor and get feedback, on which devs can make updates. Rinse and repeat
Invokes Services
Managed connectors
App Service APIs
API Management
Azure Functions
Workflow
HTTP + Swagger
HTTP
Behaviors
Retry policy
Run after
Limit
Operation Options
Response
Webhook
Delay
Terminate
Message Handling
Compose
Query
Table
Request schema
Parse JSON
Xpath & XSLT
XML Validation
Flow Control
Scope
Foreach
DoUntil
Switch
Condition
Expressions & Operators
Invokes Services
Managed connectors
App Service APIs
API Management
Azure Functions
Workflow
HTTP + Swagger
HTTP
Behaviors
Retry policy
Run after
Limit
Operation Options
Response
Webhook
Delay
Terminate
Message Handling
Compose
Query
Table
Request schema
Parse JSON
Xpath & XSLT
XML Validation
Flow Control
Scope
Foreach
DoUntil
Switch
Condition
Expressions & Operators
HIPAA - Health Insurance Portability & Accountability Act with Business Associate Agreements (BAAs).
PCI DSS - Payment Card Industry Data Security Standards Level 1 version 3.1.
SOC - Service Organization Controls standards for operational security
GDPR – General Data Protection Regulation