High Availability and Resiliency on IBM Cloud

What makes a cloud computing infrastructure resilient?

In IBM Cloud, a resilient infrastructure keeps your cloud solution running at acceptable levels, even if one or more areas of the infrastructure experiences a minor or major problem:

• Power loss
• Planned or unplanned maintenance
• Hardware failure
• Network failure
• Natural disaster
• Ransomware

Resilient solutions are made up of multiple components. Different components can satisfy your business and technical requirements:

• High Availability - System designs that provide continuous service, often by using redundant components to protect against failure. Redundancy is using multiple identical resources in different locations to protect against a single point of failure.
• Back up - Continuous backup of critical data for recovery to continue to provide service if key data is lost or inaccessible.
• Disaster Recovery - Regaining access to the infrastructure after extreme events such as natural disasters, cyberattacks, or wide range power outages.

IBM Cloud and customer responsibilities

With IBM Cloud, you have several methods to build resilient architectures to choose from:

• Standardized user interface that makes building custom infrastructures as simple as provisioning a server.
• Terraform scripts that provide a CLI to automate your deployment consistently across different projects or locations.
• Schematics workspaces that provide a user interface to customize and automate your deployment.

Regions and availability zones

IBM Cloud data centers are in different geographical areas, countries, and regions. Each region has one or more availability zones, which are specific physical locations. Data centers can be single-zone, multi-zone region, or cross-multi-zone region.

Single-zone

A singl availability zone (SAZ) provides redundant resources with low to medium cost. Use an SAZ for medium level applications with less stringent resiliency requirements.

The single zone provides multiple VSIs for web, app, and db tiers in a single zone.

Multi-zone region

Use an MZR for core business applications that have stringent resiliency requirements. The MZR scales services based on events at a medium to high cost.

An MZR provides these benefits:

• Consistent cloud services across the different zones
• Better resiliency, availability, higher interconnect speed between data centers for cloud platforms
• Infrastructure services such as IBM® Cloud Object Storage, and IBM Cloud load balancers

These features can be critical to your applications. Deploying the application in an MZR rather than a single zone can increase the availability from 3 9’s to 4 9’s when deployed over three zones.

Cross-multi-zone

Use a cross-MZR for business critical or global services that have stringent resiliency requirements. In an Active/Standby configuration, your resources are available during a regional failure and you have a way to recover services after an event at a medium to high cost. In an Active/Active configuration, you have near zero downtime and loss at a higher cost.

Cross-multi-zone provides the benefits of a multi-zone arrangement across multiple zones.

What levels of resiliency do the different zones and regions provide

Whether the data centers are single-zone or multi-zone, all of them avoid a single-point-of-failure (SPOF) between zones and regions by providing:

• Multiple power feeds
• Fiber links
• Dedicated generators
• Battery backup

While all the data centers have multiple power feeds, several of the more mature sites have some 1U single socket server chassis that might not accommodate a dual power feed. If you have a 1U single socket server in one of these sites, you might want to consider a 2U chassis with redundant power supplies. For more information about availability zones, see Locations for resource deployment.

Add on features for highly available compute resources

At a minimum, deploying more than one virtual server instance helps make your solution highly available. IBM Cloud has more features that you can add to your solution:

• Placement groups that distribute your VSIs on different hypervisors or on different hypervisors with different power sources.
• Auto scale that aoutmtically scales your number of VSIs up and down based on capacity.

Placement groups

You can use placement groups to ensure that multiple virtual server instances are on different physical servers within a data center or availability zone. Placing virtual server instances on different physical servers improves your overall redundancy and availability story. You can improve your redundancy more by spreading placement groups across different availability zones. Your business applications continue to run and provide service to your customers even if one or more virtual server instances are lost.

You have 2 different placement algorithms available to choose from:

• Hypervisor spread - Ensures the VSIs are distributed across different hypervisors. There can be up to 12 VSIs in the placement group.
• Power spread - Ensures the VSIs are placed not only different hypervisors but also the hypervisors are placed on different power and network, essentially on different racks. There can be up to 4 VSIs in the placement group.

Auto scale

You can use Auto scale to help optimize your compute costs. Auto scale adds and removes the virtual server instances dynamically to ensure you have the right amount of capacity for the current demand. You pay for what you need when you need it. The additional resources are removed when they are no longer needed.

Using LBaaS to help increase resiliency

IBM Cloud load balancer as a service (LBaaS) is a network offering that you can use to increase resiliency within a region. IBM Cloud provides multiple network devices and connections to ensure that your servers and data storage components are always in contact with each other. When you spread your workload across multiple machines, consider using a load balancer. A load balancer provides a well-known destination to your application. It also provides redundancy and scaling.

For more information about the different types of LBaaS and capabilities, see Load balancers for VPC overview. To increase resiliency across regions, you can use the global load-balancing functions of Cloud Internet Services (CIS).

Using snapshot for backups

Some occurrences or events can put your business continuity at risk such as ransomware, data loss or corruption, accidental deletion, or catastrophic events. Make backup a part of your data resiliency plan to recover from these types of events. Backup is a point-in-time copy of your data from either your boot or data volumes. For more information about using IBM Cloud snapshot backup capabilities, see Creating Snapshots.

High Availability components and tutorials

For more information about High Availability solutions, see: