Geographically distributed. Fault tolerant. Highly available. These are descriptions that surface when discussing Cloud infrastructure, but how often are they put into practice?
Truth is, the vast majority of enterprise cloud applications operate from one Amazon availability zone (AZ) or one privately owned data center. In some cases, there is operation out of a second datacenter as a cold failover.
As evidenced by recent Amazon outages, these configurations lead to performance degradation or downtime to varying degrees for many cloud-based companies.
For cloud application stacks with a high-trust requirement, such as PingOne, those scenarios are not acceptable.
The PingOne team believes customers can never waver on their trust in our systems so we have developed and architected a cloud identity management solution to exceed their expectations.
We challenge ourselves by asking many questions of which here are a few:
How can we build a multi-tenant architecture?
How can we build autonomous data center failover?
How can we achieve the flexibility needed under Brown out conditions?
How can we measure connectivity problems to trigger failover?
How do we keep our data consistent across multiple geographic locations?
For some service providers, rather than answering these questions adequately it is far easier to accept downtime or performance degradation burps, and take the lump on the head from the customer.
But likely that’s not an option for you? What if your requirements won’t allow it? What if your company won’t allow it? And surely, your customer's won't allow it?
At Ping Identity, we set a high bar. It's how we operate. It's what we believe. Internally, we think that's what being "Ping" is all about.
For PingOne, that spirit lives in our Site Reliability Team.
Over the next several articles, I’ll talk about the linchpin in our geographically distributed enterprise application, The Data Layer.
Teaser: Next article I’ll provide a broad overview of our data layer, what technologies we use and why we use them.
About the Author:Michael Ward is a Site Reliability Engineer for Ping Identity, specializing in high availability solutions and system architecture. firstname.lastname@example.org