TMCnet Feature

The challenge of how to scale ADC (News - Alert) solutions should be addressed from the bottom-up, beginning with a robust foundation of hardware and software architecture that allows highly scalable ADC services inside the “borders” of the physical platform.  

In order to scale beyond the physical scope, today’s ADC solutions support virtualization capabilities that allow it to reside as a virtual appliance on mutli-purpose servers and use existing data center resources. Moreover, by intelligently managing the physical and virtual resources via a “shared resource pool” concept, it is possible to dynamically optimize the overall ADC solution resources to fit the most demanding performance requirements holistically, thus improving the overall efficiency of the data center.

Additionally, scaling out beyond the data center network borders into the cloud requires the ADC solution to be well integrated into the cloud’s ecosystems. Once this challenging task is overcome, a smooth operation is needed to transfer the required ADC resources, on-demand, to the cloud and manage it efficiently within the cloud. 

Let’s take a look at the five ADC scalability leverage points, which are summarized as follows:

1.       Strong, scalable foundation platforms

An ADC software architecture which enables an “on-demand” business model can help companies to scale only when needed, with no impact on their business operations. It begins with the ADC hardware platform shipped with very strong computing power capable of serving a mutli-gigabit data center environment, but can limit its processing power per license. On- demand ADC software architecture allows a company to begin with low-end processing (e.g., 100-200 Mbps) and scale up only when needed by purchasing a higher throughput license with no hardware replacement and no system reboot. 

2.       Leveraging virtualization technologies

To further leverage the on-demand model, a modern ADC solution provides an architecture that provides mutli-instance ADCs on the same platform. This ADC virtualization capability supports data center consolidations. Instead of using multiple ADC boxes to load balance different applications, one box may be used with multiple ADC instances, each instance serving a different application with completely isolated resources. This capability ensures performance predictability and application resilience, allows optimizing ADC resources and the use of more boxes only when they are truly needed, and most importantly, enables full utilization of these resources for improved IT efficiency.

3.       Dynamic resource pool management

ADC virtualization technologies allow IT managers to create, remove, migrate and change the characteristics of ADC instances, either on a physical ADC appliance, or on controlling virtual ADC appliances. An ADC resource pool management capability that is integrated into the data center orchestration system can manage all the data center elements and allows a further scaling of the ADC solution by dynamically allocating the ADC resources per need without investing in new hardware. This reduces operational complexity. For example, application bursts from sudden usage demands sometime require a temporary “boost” of ADC power that can be handled seamlessly by the pool management system. At the same time, available computing resources would be allocated to other services or applications in order to guarantee their SLA - with no human intervention.

4.       Scaling out to the cloud  

Cloud service providers require a scalable ADC solution which is aligned with the provider’s tiered approach to service level guarantees. ADC solutions for a cloud provider include powerful ADC appliances with mutli-instance capabilities that can serve hundreds of customer on the platform, while maintaining full isolation. The ADC resource pool manager, working in concert with the cloud ecosystem, offers an automated workflow for adding services for new customers based on the required SLA. Customers with a low SLA (e.g., best-effort based service) will be automatically assigned to a soft ADC resource, while customers with high SLAs will be assigned to a more predictable and guaranteed resource on a dedicated ADC appliance.  

5.       The vision: An open-flow, new way to scale ADC solutions

The computer networking market has been evolving into a new era that may use existing router and switch infrastructures as “enforcers” of some ADC functions. The ADC intelligence will be placed in a control plane that provides instructions to this network equipment. This direction would, of course, allow an infinite pool of ADC enforcers, but much more needs to occur before we see it implemented.  

Combined, these five points are a helpful guide to addressing the best approach to leveraging a scalable ADC solution from the bottom up. Implementing these points can set you on the right path to optimize ADC resources in order to meet performance and SLA demands and add a new level of efficiency in the data center.

Avi Chesla is Chief Technology Officer at Radware (News - Alert). To read more of his articles, please visit his columnist page.