Use Case 2 - Solution
Use Case 2 - Solution
Use Case 2 - Solution
Introduction
I. Purpose
• This document captures the current architectural components and configurations of Cloud components, deployment
architecture and migration plan for applications
2. Problem statement
• Issues with handling the traffic spikes during yearly sales
• Lease for 2 Data Centres is expiring soon. Need to find a secondary infrastructure to host the applications.
• Issues with releasing new features. Low frequency of deployments.
• Overhead on the operations team for managing the infrastructure.
3. Observations
• Infrastructure Auto scaling is required to address issues arising from traffic spikes
• Implement a zero-downtime deployment strategy to reduce time to market and release new features frequently
• Need to move away from traditional way of managing infrastructure to IAAS/PAAS
• Database should be deployed within India only
4. Assumptions
• It is mentioned that customer has containerized few of the applications, but not specified which one in particular. It
is assumed that the web tier (Nginx) and application tier (websphere) both are containerized.
• The bandwidth requirement for the internal and communication among the servers is not mentioned. Also, cost does
not seem to be a major concern here as it is not mentioned. 150 servers that need to talk to each other. The
requirement is assumed to be around 5GB-10GB between the data centres and the cloud
• Assumed that all 3 data centres host 50 VMs each on which load is almost equally distributed.
• Location and HA setup of oracle database is not mentioned. Assumed to be present in DC-3 and DC-2 with Data-
Guard enabled
5. Key Decisions
• VMs from the 2 data centres for which lease is expiring, will be decommissioned. The 3rd data centre for which there
is 2 year lease period remaining, will continue to host the applications as the cost is already paid.
• Since application running on websphere is already containerized, we can use GKE service to host these applications.
It will give us additional features of auto-scaling and descaling as per demand
• The network bandwidth requirement between the DC-3 and the GCP is assumed to be 5GB-10GB. Google cloud
interconnect will be used to communicate between the DC-3 and GCP VMs/services.
• Region ASIA-SOUTH1 is used so that data resides within India.
• ANTHOS will be used to manage the resources across on premises (DC-3) and GCP
Data Center - 3 PROD : Cloud Architecture
VPC network
On-Premises
Subnet Google Cloud region
asia-south1
VPC Subnet
Oracle Oracle
Database Database
Server (Slave) Server (master)
Data Guard
DC-3 to DC-2
CDC
Change Log Collector
zone-a
Oracle Oracle
Cloud SQL
Database Database
Master
Server (Slave) Server (master)
Data Guard
DC-3 to DC-2
CDC
Change Log Collector
zone-a zone-b
Oracle Oracle
Cloud SQL Cloud SQL
Database Database
Master Replica
Server (Slave) Server (master)
Data
Data Guard Replication
DC-3 to DC-2
zone-a zone-b
Data
Replication
Anthos
On-premises network
GCP
Dashboard
Cloud Run
Web-Tier
(Nginx) On-Prem
Cloud Run
on Cloud Run
Cloud Run
GKE-Cluster
App Tier
On-Prem
(WebSphere) GKE Cluster
GKE
on GKE
Developer End-User
New commit pushed to repo Build code, Push docker image too GCR Deploy to GKE
Build artifacts - image