This article is for IT infrastructure architects, system administrators, SAP Basis architects/ administrators who want to implement or extend SAP HANA workloads on AWS.
If there is an active AWS account already and you are familiar with AWS services and SAP HANA, then better to launch the Quick Start for building a scenario in a new or existing virtual private cloud (VPC). Each deployment takes less than one hour to complete, depending on the selected configuration.
In the background, Quick Start uses AWS CloudFormation, the AWS Command Line Interface (AWS CLI) for Linux, and custom scripts to deploy SAP HANA on AWS. AWS CloudFormation provides an easy way to create and manage a collection of related AWS resources, provisioning and updating them in an orderly and predictable fashion – which minimize the unwanted issues during deployment.
While using Quick Start, one can opt for the following – which is optional in nature:
automated installation of SAP HANA software.
Another EC2 instance with Windows Server in the public subnet to host SAP HANA Studio to administer SAP HANA database. One can install SAP HANA Studio manually.
The Quick Start follows security best practices by establishing a VPC with public and private subnets. For secure access, the SAP HANA server is placed in the private subnet, which is not directly accessible from the internet. One can access to SAP HANA server either through AWS Systems Manager Session Manager, the bastion host, or by using the optional Windows Server instance with an SSH client.
The Quick Start provides the following deployment scenarios:
Single-AZ deployment (HANA Standalone/single-node and HANA Scale-out/multi-nodes)
Multi-AZ deployment (HANA Standalone/single-node)
Deployment scenario #1:
Single Node SAP HANA (HANA Standalone) within Single AWS Availability Zone
This scenario uses a single AWS Availability Zone for the deployment. The SAP HANA single-node option provisions a single EC2 instance for SAP HANA in the private subnet of the Availability Zone.
Architecture for Single Node SAP HANA (HANA Standalone) within
Single AWS Availability Zone
Deployment scenario #2:
Single Node SAP HANA (HANA Standalone) within Multiple AWS Availability Zones
This scenario uses two AWS Availability Zones for the SAP HANA environment. It provisions two SAP HANA servers (primary and secondary) in separate private subnets in those zones and configures them for high availability. This scenario is popular with those – who are looking for the high availability options.
Architecture for Single Node SAP HANA (HANA Standalone) within
Multiple AWS Availability Zones
For this architecture, one must opt for OS Cluster from the following available choices:
- SLES High Availability Extension (HAE) or
- RHEL for SAP High Availability solutions.
Generally, SAP HANA high availability deployment supports Performance-optimized and Cost-optimized options.
AWS Quick Start supports only Performance-optimized option. In this option, SAP HANA System Replication (HSR) is leveraged for data replication between primary & secondary HANA nodes.
In Cost-optimized option, the secondary node can be used for two systems (but single system will be always active at a single point in time).
Deployment scenario #3:
Multi Node SAP HANA (HANA Scale-out) within Single AWS Availability Zone
This scenario uses a single AWS Availability Zone for the deployment. The SAP HANA Scale-out or multi-node option provisions upto five EC2 instances for SAP HANA in the private subnet of the Availability Zone.
If there is a business requirement of deploying SAP HANA scale-out clusters that are larger than five nodes, then contact AWS via e-mail to saphana@amazon.com. In that case, a customized AWS CloudFormation template will be provided by AWS for deploying SAP HANA clusters.
Architecture for Multi Node SAP HANA (HANA Scale-out) within
Single AWS Availability Zone
Additionally, the security group or firewall rules need to be extended to allow for inter-node communication between all SAP HANA servers over the appropriate ports and protocols.
Picture courtesy : AWS
Comments