Fundamentals
Cloud concepts
Azure core services
Compute
Networking
Data & Storage Services
Other Services
Identity
Compliance
Security
Cost
Core Cloud Services
Compute
Storage
Networking
Application Services
Analytics
Scalability
Increase or decrease resource based on workload demand
Vertical Scaling : also known as scaling up
Add additional resources to increase the power of the workload
Eg: Add additional CPUs to a virtual machine
Horizontal Scaling : Also known as scaling out.
Elastic
Major pattern which benefits from cloud computing
As your workload changes, resources can be changed to compensate ( Up or down )
Example : Seasonal demand for retail web site.
Static Scaling 2. Elastic Scaling
What is an SLA ?
A service level agreement ( SLA ) is an agreement with the business and application teams on the expected performance and availability of a specific service.
Define SLA’s for each workload
Dependency mapping : Make sure to include internal/external dependencies Identify single points if failure.
Example : workload requires 99.9% but depends on a service that is only 99.9%
Key terms :
Mean Time to Recovery ( MTTR ) : Average time to recover service from an outage.
Mean Time Between Failures ( MTBF ) : Average time b/w outages
Recovery Point Objective ( RPO ) : Interval of time in which data could be lost during a recovery.
Eg: 5 Minute RPO means up to 5 minutes of data could be lost.
Recovery Time Objective ( RTO ) : Time requirement for recovery to be completed in before there is business unit.
Disaster recovery
Planning for catastrophic failure of workload
Region to region failover
On-premises to cloud failover
Automation and orchestration
Fault Tolerance
Redundancy is built into services so that if one component fails, another takes its place.
Reduces impact when disasters occur.
HA Examples
Host Outage
When an underlying host has a catastrophic failure, the virtual machine will automatically be restarted on another host.
Availability sets and zones further increase the availability.
Cross region deployment
An application is deployed in a configuration to be highly available across regions.
When a service in one region has an outage, traffic can continue to run in the second region.
Cloud Service Models
Traditional
Infrastructure ( as a service )
Platform ( as a service )
Software ( as a service )
Cloud Economics
Cloud benefits
Cloud providers can pass on economies of scale to consumers
Acquire hardware at Lower costs
Local government deals
Datacenter efficiencies
Capex ( Capital Expenditure )
Spending on infrastructure is completed upfront
Cost written off over period of time
Operational Expenditure ( OpEx )
No Up-front cost
Pay for service as you consume it
Deduct from tax bill in same year as expense occurs
Capex costs
Server costs
Storage costs
Network costs
Backup and archive costs
Datacenter costs ( including DR ).
Opex costs
Server lease costs
Software and feature leases
Usage/demand cost scaling
Azure Resource Manager ( ARM )
Resource : Individual manageable item available to you in azure
Resource group : container where you can house your resource for management
Resource provider : provider of services you can deploy in azure ( Microsoft compute ).
ARM templates : files used to define resource you wish to deploy to a resource group.
ARM Templates overview
Apply infrastructure as a code
Download templates from azure portal
Author new templates
Use Quickstart templates , provided by Microsoft
ARM File Types
ARM Template File : Describe the configuration of your infrastructure via a JSON file
ARM Template Parameter File : Separate your parameter ( optional )
Deployment scripts : E.g : Powershell for deployment
ARM Template Constructs
Parameters
Variables
Resources
Outputs
Azure virtual machine
VMTYPES
A - Basic : Basic version of the A series for testing and development
A - Standard : General-purpose VMs.
B - Burstable : Burstable instances that can burst to the full capacity of the CPU when needed.
D - General purpose : Built for enterprise applications. DS instances offer premium storage.
E - Memory optimized : High memory-to-cpu core ratio. ES instances offer premium storage.
F - CPU optimized : High CPU core-to-memory ratio. FS instances offer premium storage.
G - Godzilla : very large instances ideal for large databases and big data use cases
H - High performance compute : high performance compute instances aimed at very high-end computational needs such as molecular modeling and other scientific applications.
L - Storage optimize : storage optimized instances which offer a higher disk throughput an IO
M - lARGE memory : another large scale memory option that allows for up to 3.5 TB of RAM
N - gpu ENABLED : cpu-enabled instances
SAP HANA on azure certified instances : Specialized instances purposely built and certified for running SAP HANA.
VM Specializations
S : premium storage options available ( Eg: DSv2 )
M : larger memory configuration of instance type ( Example : Standard A2m_v2 )
R : Supports remote direct memory ( RDMA ) Eg: H16 mr
Module VM Availability :
Potential for VM impact :
Planned maintenance
Unplanned hardware maintainence
Unexpected downtime
Availability Sets
Group two or more machines in a set
Separated based in Fault Domains and Update Domains
Fault Domains and Update Domains
Planning for availability
1. Web tier availability set
2. App tier availability set
3. Data tier availability set
App Services
Introduction to WebApps
- Web Apps
- Mobile Apps
- Logic Apps
- API Apps
App Service Environments ( ASEs )
- Fully isolated environment
- For high-performing apps - high CPU and/or memory
- Individual or multiple service plans
- 2 ways to deploy : internal or external
- Created in a subnet via a Vnet which achieves a isolation
- Note : May take a few hours to spin up
Compute services - containers
Containers :
Standadized packaging for software and dependencies
A way to isolate apps from each other
Works with linux and windows servers
Allows separate apps to share the same OS kernel
Application Modernization
Monolithic App Issues
- Minor code changes required full recompile and testing
- Application becomes a single point of failure
- Application is difficult and often expensive to scale
Microservices:
Break application out into separate services
12-factor apps
- Make the app independently scalable , stateless, highly availably be design
Comparing Monolithic and Microservices
Monolithic : Simple deployments inter-module refactoring, vertical scaling , technology monoculture
Microservices : partial deployments strong module boundaries horizontal scaling technology diversity
Serverless Computing
What is serverless computing ?
Fully managed services
Only pay for what you use
Flexibility to scale, as needed
Stitch together applications and services seamlessly
Azure serverless computing services
Azure functions
Logic Apps
Event Grid
Azure functions key features
- Program languages C#, F#, JavaScript, Java ( Preview )
- Pay-per-use pricing A. Consumption Plan B. App Service Plan ( Run on the same plan as other services ).
- Integrated security with OAUTH providers ( Azure AD , Facebook etc ).
- Code in the portal or deploy via DevOps tools.
Logic Apps Key Features
- Workflow Engine
- Used to orchestrate and stitch together functions and services ( Just like regular orchestration tools ).
- Visualize , Design , Build , Automate
Logic Apps - Key Constructs
Triggers ——————————> Action
VNET
Core VNET Capabilities
- Isolation
- Internet Access
- Azure resources ( VMs and Cloud Services )
- VNet Connectivity
- Onpremises Connectivity
- Traffic Filter
- Routing
- Primary building block for Azure Networking
- Private Network in Azure based on an address space prefix
- Create subnets in your VNet with your own IP ranges
- Bring your own DNS or Use azure provided DNS
- Choose to connect the network to on-premises or the internet
Hybrid Connectivity
Hybrid Connectivity Options
Site-to-Site ( S2S )
ExpressRoute
Point-to-Site ( P2S )
S2SVPN gateway connection is a connection over IPsec/IKE ( IKEv1 or IKEv2 ) VPN tunnel.
Requires a VPN device in enterprise datacenter that has a public IP address assigned to it.
Must not be located behind a NAT
S2S connections can be used for cross-premises and hybrid configurations
P2S
Secure Connection from an individual computer. Great for remote worker situations.
No need for a VPN device or public IP. Connect wherever user has internet connection.
OS support: Windows 7,8,8.1 ( 32 and 64 bit ), Windows s10, Windows Server 2008 R2, 2012, 2012 R2 64-bit
Throughput up to 100 Mbps ( unpredictable due to internet ).
Doesn’t’t scale easily, so only useful for a few workstations.
VPN Gateway SKUs
