Today, organizations need to support all types of end users that access services and applications in the data center. This means longer and more complicated application deployments on the ADC.
To successfully deploy an application on the ADC, one needs a lot of specific knowledge about the application. For example, you need to know what health checks to do, what should be synched to the backup device, acceleration enhancements and more. The applications require a significant testing phase and there are many parameters on the ADC that need to be monitored in the day-to-day operations.
The other challenge in application deployment is related to the ADC deployment models. Until today, organizations adopted two ADC deployment models: the dedicated ADC model where one physical ADC serves a single application, and the shared ADC model, where a single physical ADC serves multiple applications. The problem with the dedicated model is that it is expensive; there are high costs associated with ADC hardware and high operational costs of power, cooling and rack-space. In addition, this solution is not scalable, so the CFO will not authorize funding to buy a new ADC for every new application.
The shared ADC model also has its challenges, like resource sharing between all the applications potentially leading to unreliable performance. A problem in one service might affect the entire device. When applying a software upgrade to the ADC it impacts all the applications on the device and requires a maintenance window. Since there is one big configuration file, troubleshooting is also more complex.
Instead, there’s a better way to deploy an application in the data center.
To overcome the first challenge of long deployment time of new applications, the IT team must have more than the traditional network-centric view, while also automating details of the configuration and ongoing configuration synchronization, an application-specific view. The toolkit should include automation technology to manage the application delivery services from an application-centric view. The complete tool will offer three components: configuration templates and wizards to help with application rollout, application centric screens that show only the relevant application information to simplify day-to-day operations, and it should provide in-depth reporting and analytics information about the application, its performance, logging info and other aspects of the network. The result is a much faster application rollout, optimized application delivery on the ADC and centralized application operations.
To overcome the challenges of traditional ADC deployment models, a new standard for ADC deployment in the data center should be considered: The creation of a virtual application delivery infrastructure that allows agility and scaling in concert with the server resources, while maintaining the highest level of security and can even reduce cost. The fabric is a combination of the abstraction of resources to create the hypervisor, virtual ADC instances, a combination of form factor of soft, hard and vADC’s as well as a stocked toolkit and integration into the data centers virtual ecosystem.
This is the foundation that starts with the ADC hypervisor, which can run multiple virtual ADC instances. Consolidate the dedicated ADCs and create a vADC for each one of the ADCs that has been consolidated. Then replace the shared ADC device and create a vADC per each one of the applications in the shared device.
In my next post, I’ll further discuss the vADC Application and shaping tools for ADC deployments. Until then, be sure to leave any comments with questions and or additional insights.