Cloud Networking Pre-requisites
- All the networks created by us in any cloud are private by default.
- Enabling public ip to the virtual machine
- CSP will have pool of public ip addresses out of which one ip is selected randomly and assigned to your network interface.
- When the vm is shutdown, the public ip will be released and it will go back to the pool.
- If we want the public ip to be same even after stopping and starting the vm, we need to make static public ip reservations.
Auto Scaling (Elasticity)
- When the number of servers hosting the customer facing applications over internet are scaled, each server gets its own private ip. Public ip also can be allocated
- Questions:
- Do we really need a public ip for each vm
- What is that public which customers are going to use to access the application.
- OSI Model of Networking
- In the above case, the same application is scaled so we can use a Load Balancer which is Layer 4 aware
- Load Balancer performs
- Health checks/probes to ensure whether the service is up or not.
- For Healthy servers, the traffic will be forwarded in Round Robin Fashion
- Now, The vm’s can be private and its enough if the load balancer has public ip.
- Layer 7:
Hybrid Networking
- Site-to-Site VPN
- Next Steps:
- Multiples sites to a network vpn connection.
Disaster Recovery and Multi Region Setups
-
Active-Passive:
-
Active-Active: