Every business has specific requirements with regards to the availability (uptime) of their applications. Digital transformation is gradually evolving applications of nearly every type of business into business-critical applications. When applications are unavailable, most companies are unable to run their business. While ‘born in the cloud’ applications, called Cloud Native Apps, are self-proficient in meeting uptime requirements and don’t rely all that much on the underlying IT infrastructure, traditional client/server or so called n-tier applications rely heavily on availability and recoverability measures in the infrastructure layer. Creating a highly resilient IT infrastructure across multiple physical locations to run these traditional applications is still the best, and often only, way to guarantee application uptime. To successfully architect a highly resilient IT infrastructure, you can make use of the concept of Regions and Availability Zones.
What are Regions and Availability Zones
Ever since AWS started delivering their cloud services out of different physical locations that were hierarchically and logically divided in so called Regions and Availability Zones, lots of customers and vendors have embraced these constructs. To clarify, a Region is literally a separate geographic region (for example US East Coast and US West Coast). Each Region contains two or more isolated physical locations or sites called Availability Zones. An Availability Zone is logically speaking a single isolated fault domain and (most times) practically speaking a single physical data center or site. An Availability Zone should be physically isolated from all other Availability Zones within the Region. A single power outage for example shouldn’t impact more than one Availability Zone.
Why are customers using Regions and Availability Zones
The benefit of dividing geographically separated data center locations into these logical constructs is that customers are able to implement specific availability and recoverability measures at both the Availability Zone level and at the Region level. From an infrastructure perspective, two Availability Zones within a single Region could be an ideal use case for a metro cluster with a synchronously replicated storage solution or a stretched storage solution such as VMware vSAN.
This will provide very fast and fully automated recovery of workloads with zero data loss in case of an Availability Zone failure. A stretched cluster solution, such as VMware vSAN Stretched Cluster, writes data to each site in a synchronous fashion to guarantee zero data loss in case of a host or site failure.
At the Region level, network constraints due to geographical distance induced latency are highly likely to prohibit the use of synchronous storage replication or a stretched storage solution. VMware vSAN for example supports a Round Trip Time (RTT) of 5 ms maximum for stretched clustering. RTT is the duration in milliseconds (ms) it takes for a network request to go from a starting point to a destination and back again to the starting point. If the RTT is too high, performance will suffer greatly.
As an alternative, asynchronous storage replication might be considered. You can specify an interval in which data changes in the primary site are replicated to the recovery site. If you specify an interval of 1 hour and a site failure occurs, worst case, at 59 minutes and 59 seconds and 999 ms, you will lose all changes made to the primary site in that period of time. With asynchronous replication you should carefully align your synchronization interval to the specified Recovery Time Objective (RPO) requirement.
VMware Validated Design for Software-Defined Data Center (SDDC)
After revolutionizing the x86 server virtualization market with VMware vSphere, VMware introduced a vision for a Software-Defined Data Center. In an SDDC, all data center resources such as compute, storage and networking are abstracted in virtualized pools of resources and consumed through extensive automation. This means that delivery of data center resources such as Virtual Machines, storage and security services can be done on-demand using self-service provisioning. To provide optimal guidance during the design and deployment of an SDDC, VMware created a reference architecture with accompanying operational documents called VMware Validated Design (VVD) for SDDC. This is basically an architecture blueprint on how to deploy, operationalize and manage a VMware SDDC according to VMware best practice guidelines.
The VVD for SDDC also uses Regions and Availability Zones, VMware is now also actively using them in commercial offerings that are based on the VVD for SDDC reference architecture, such as VMware Cloud Foundation and VMware Cloud on AWS:
The VMware Validated Design for SDDC R4.2, which was released in February 2018, first introduced a Multiple Availability Zone architecture leveraging VMware vSAN in a stretched cluster configuration as a disaster avoidance measure. This allows customers to run their workloads seamlessly across sites and guarantee an RPO of near zero in case of an Availability Zone level failure.
Customers can further improve the recoverability and general availability of their environment by adding additional recoverability measures at the Region level. The VVD for SDDC describes the use of asynchronous storage replication of Virtual Machines at the server virtualization layer, and the use of specific disaster recovery software to initiate a fully orchestrated failover of Virtual Machines in case of a full Region failure.
The dual Region setup of the VVD for SDDC also allows customers to configure cross-region replicated backups. This ensures backup data is safely stored in a separate geographic location.
Vembu BDR Suite provides business continuity across Regions and Availability Zones
Vembu BDR Suite provides the following features to provide business continuity across Regions and Availability Zones:
- Host Level VMware Backups protect vSphere and vCenter environments using the VMware vStorage APIs (VADP)
- Offsite Disaster Recovery provides extended data protection to critical business data by replicating backup data to an offsite data center
- VM Replication provides replication technology to periodically replicate virtual machines from one ESXi host to another ESXi host
In my following blog, I will provide a step-by-step configuration guide to provide specific details on implementing these three Vembu features in a Dual Region, Multiple Availability Zone VMware Software-Defined Data Center.
Follow our Twitter and Facebook feeds for new releases, updates, insightful posts and more.
