In every data center, there are applications that are essentially treated like they are just not important enough to warrant the investment it takes to provide reliability, security and performance-related services. The problem is, they are important.
If the organization made the investment to have the application developed or to purchase it, it is important. It serves a business purpose and has a reason for being. It should be afforded the benefits of application services.
The unfortunate reality is that providing every application with the application services they need is costly. Traditionally, application service delivery models are based on high-availability principles that rely on redundancy and isolation to ensure availability, security and performance.
This makes it difficult to justify the expense for each and every deployed application and is particularly true for newer applications that have not proved their business value yet. But it’s a chicken and egg story, without data to prove value, organizations take a “wait and see” attitude.
Hence, an application that rapidly gains adoption and usage might warrant some services. However, without those services the application is more likely to face challenges to adoption and usage. The patient consumer is a fantasy. Consumers are quick to abandon poorly performing or insecure applications.
There must be a better way of getting this done. F5 proposes a high performance services fabric that can change the economies of scale associated with delivering application services. It takes a leaf out of the cloud computing book. Economies of scale in cloud are achieved through abstraction of resources.
Compute, network and storage resources are abstracted and pooled together such that they could be provisioned as services on-demand. Economies of scale ensure that the cost of using those services decreases, making them affordable for even the smallest of organizations.
In the data center, however, similar economies of scale has been difficult to achieve because the abstraction at the network layers has remained elusive. SDN has recently emerged as a front-runner in the data center as a provider of that abstraction, turning disparate network elements into a cohesive fabric of network resources, dynamically adjusted to deliver the best performance and availability to any application that might be delivered over formerly rigid pipes.
But SDN doesn’t provide an easy answer for layer 4-7 (application) services. Application services must be able to match the economies of scale offered by server virtualization and SDN/NFV if economies of scale in the manner of cloud computing is to be realized.
By applying the principles of traditional high-availability architectures to clusters of service platforms, along with a multi-tenant and application-centric approach to scalability and reliability, organizations will find a significantly more flexible and cost-effective means of deploying application services to every application, not just those deemed mission critical.