How to Converse In Cloud: Reaching Multicloud High Availability

Multicloud high availability (HA) can help protect your enterprise’s valuable application and data assets against costly system and power outages and ensure business continuity (BC) and disaster recovery (DR). But how do you best protect against planned and unplanned downtime for both on-premises and cloud infrastructures? Because there are so many choices of HA scenarios using multicloud architectures, the answer is “It depends.” Let’s look at a few of the more common use cases.

Leverage private interconnection to the cloud

I’ll be focusing on three different levels of high availability in the cloud in this article: data-to-data, application-to-application and cloud-to-cloud. But first, let’s look at how private interconnection plays a critical role in ensuring the direct and secure connectivity to cloud services. Dedicated, private interconnection between on-premises or colocated IT infrastructures and public clouds are available from practically every leading public cloud provider, including:

These and many more cloud direct connects are all supported on Equinix Cloud Exchange Fabric™ (ECX Fabric™) to deploy hybrid/multicloud infrastructures on Platform Equinix® using software-defined interconnection. I’ll be talking more about how this is done in the examples below.

Start with some redundancy

Replicating your IT infrastructure, including cloud-based services, in regionally distributed data centers is key to laying the foundation for multicloud high availability. Do you need a robust DR infrastructure in a different geographic location from your primary production site? A high-performance data backup and recovery environment that meets security policies and compliance regulations? Or an HA architecture that ensures reliable application failover during unplanned downtime or planned maintenance? Then mirroring your primary IT infrastructure to a secondary one, either on-premises or in the cloud, is an ounce of prevention worth investing in. Selecting a global interconnection and data center platform, such as Equinix, helps you to reduce your overhead in both HA CAPEX and OPEX. It also ensures consistency in terms of the overall data center operations across multiple data centers around the world.

You could also opt to duplicate your entire IT infrastructure in the cloud on Platform Equinix, which would really reduce your CAPEX costs by totally converting them to a usage-based OPEX model. And with today’s virtualization technologies, you don’t have to completely duplicate your primary infrastructure for quick failover and business continuity. You can dynamically launch a second instance of any mission-critical application environment in the cloud.

Think about the type of high availability you need

Before you set off replicating systems, storage and applications on multiple cloud platforms, think of where and how you would use high availability. What data center infrastructures or application and data environments do you want to protect the most? Where do you want to set up your primary production and secondary recovery zones (either on-premises or in the cloud) to best protect data and applications? These are all questions that need to be answered before you can decide where and how to deploy a multicloud HA infrastructure. Let’s see what’s possible in the following HA multicloud scenarios:

Data-to-Data

The diagram below illustrates a use case where data protection needs to be in a more controlled, private, on-premises environment for corporate security policies and compliance regulations, however, all other applications accessing the data need the compute power and scalability of the cloud. By placing the data in two geographically separate Equinix International Business Exchange™ (IBX®) data centers and providing redundant, private interconnection via ECX Fabric to both data stores between Oracle® Cloud Infrastructure (primary) and another cloud platform (secondary), then both the data and application environments are protected. The primary data store can be replicated to the secondary one between the IBX data centers and all other applications can be backed up between the Oracle Cloud and any other cloud platform for HA. This “cloud adjacent” architecture keeps data and applications accessible over the redundant software-defined and physical connections that are established between the ECX Fabric, the cloud platform’s direct connection, and your network carrier of choice.

Application-to-Application

In this example shown below, the New York IBX data center is the primary production location and the Washington IBX facility is the secondary backup site. Both are interconnected via the ECX Fabric to VMware Cloud in AWS, where application development and testing are being performed. As new applications and/or updates are ready to go into production, the user traffic can be temporarily routed to the secondary data center location in Washington, D.C., where all of the applications and data are replicated. This will ensure that there are no issues during the introduction of the new/updated application to the production site that could disrupt service to users.

Once the new application is safely running in the primary production data center in New York, then the user traffic can be seamlessly moved back to that site. The secondary site would then be updated to include the new/updated application. You can also move the user application traffic onto a multicloud environment that replicates the production environment during planned maintenance, however, it may be more cost-effective to route the application workloads and data over to a secondary data center site, than incur data egress charges from the cloud.

 

Cloud-to-Cloud

Increasingly, businesses are finding that they have to manage two cloud providers, either due to a merger or acquisition, or because their applications are more suited to one cloud platform than another. This is a perfect opportunity to leverage high availability in a multicloud environment. ECX Fabric provides software-defined interconnection between multiple cloud providers and your enterprise IT infrastructure. By replicating mission critical data and business applications across two different cloud providers with HA and data replication, then you have created a robust and reliable multicloud BC and DR infrastructure that can withstand the majority of planned and unplanned outages.

This configuration offers both hybrid cloud performance and security, and flexible multicloud high availability. IT organizations can configure on-premises applications and databases to access the specific private and public cloud platform services that best meet their needs (in this case either Microsoft Azure Stack or AWS), while at the same time gain the ability to replicate cloud-based applications and data between two different public cloud platforms where required.

These are just a few of the endless possibilities for implementing high availability using multicloud infrastructures. When deploying these HA environments on Platform Equinix with ECX Fabric, you can also create redundant network connections (both virtual and physical) using a choice of more than 1,800 network providers. This is important since you should not skimp on redundant links or network providers, as connections can get overloaded or fail during a disaster. When used in conjunction with Network Edge services, you can use the ECX Fabric Portal to spin up redundant virtual network connections to cloud platforms from providers such as Cisco, Juniper Networks and Palo Alto Networks in minutes. You can also choose from 2,900+ cloud and IT service providers in Equinix’s global cloud ecosystem.

There are many more ways to leverage the high availability capabilities of hybrid and multicloud infrastructures. The Hybrid Multicloud Blueprint can give you more ideas about how to design these and other cloud deployment scenarios.

You can also learn more about how to deploy your own customized hybrid/multicloud infrastructures in our Equinix Professional Services interactive Two-Day Cloud Readiness Workshop.

Print Friendly


Related Content