Srinivas Kotamraju is Director of Product Management for Cisco.
It wasn’t long ago that large corporations had a single data center supporting their entire business. But as modern enterprises have grown and expanded their footprint, it’s now common for them to have multiple data centers around the world.
This has created significant fragmentation and inefficiency—of technologies, processes, and capabilities. Infrastructure management, application policies, and data security are often inconsistent from site to site. And that can bring about an array of operational, governance, and compliance problems.
How does a large enterprise manage a single application with multiple instances spread across these disparate sites? How does a global service provider deliver uniformity—of application performance, availability, security, and access—to tenants that have an equally distributed footprint? And if there is a network failure or if an application goes down, how do enterprises troubleshoot and recover when multiple fabrics and multiple data centers are involved?
There are no easy answers when each data center is managed as a distinct island.
What’s needed is flexibility and simplicity to interconnect these geographically separated data centers through a single, consistent policy-driven framework. One that delivers application deployment on a global scale with fault isolation at a localized level. One that enables monitoring and troubleshooting for multiple tenants and applications spread across multiple environments. And one that provides granular testing, verification, and policy changes.
Uniformity of Policies and Processes
Fortunately, networks in multiple data centers can now be managed as a single, software-defined, application-centric network fabric. In doing so, policies, workloads, applications, services, and tenants can be deployed across—or moved among—those environments, without network-imposed constraints on latency.
Instead of managing each environment, each component, and each application individually, everything is managed via policy. Policies can easily be defined or augmented from a centralized controller and pushed to all of the associated data centers.
This creates uniformity of policies, processes, and capabilities. It reduces redundant, error prone manual tasks as well as the time, cost, and inefficiency of coordinating those tasks across geographically dispersed teams that often speak different languages. And it dramatically increases operational speed and agility, allowing application and policy adjustments to be made on the fly—on a global scale.
Scaling an infrastructure to support millions of endpoints, bringing new data centers online, and bolstering disaster recovery resources also become much faster and more efficient. Application policies and even entire data centers can be mirrored with a few clicks and pushed out to new or expanded environments just as easily.
Better Visibility, Fault Isolation, and Troubleshooting
While these network fabrics align policies and management across multiple data centers, they also boost availability via fault isolation. Each data center remains an isolated “availability zone,” where the control plane, data plane, and policy plane remain localized. This ensures there is no single point of failure, and if one data center or a localized resource in the data center goes down, the others aren’t impacted.
There is end-to-end visibility, however, which improves application monitoring and troubleshooting across sites. Modern network fabrics can provide real-time health scores of inter-site applications and tenants. If there is a problem, the fabric automatically identifies the source of the fault, with a clear description and recommended action. Advanced capabilities allow measurement of latency between endpoints to rapidly analyze performance issues.
These developments are not entirely new, of course. Data center management has been vastly improved in recent years, thanks in part to software-defined, policy-based network fabrics. And now there is a need to extend those efficiency, consistency, and simplicity gains across multiple data centers.
Why manage a collection of disparate islands when you can manage a nation of aligned states?
Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.