ADC/Load Balancer Tuning

Load balancers were created to distribute web requests across a pool of servers in order to scale processing efficiently. Over time, load balancers evolved into Application Delivery Controllers (ADCs) as they took on additional features, such as TCP optimization, SSL termination, and HTTP caching. But these complex infrastructure components come with a set of unique challenges:

• Complex configuration settings – The latest generation of ADCs are powerful solutions that come with numerous settings and knobs for tuning. The ideal settings are highly dependent on your specific network and application mix. If incorrectly configured, settings such as buffer size or the use of various TCP options can impact performance negatively and lead to lost productivity and revenue.
• Blurred lines between the application and network layers – With ADCs such as Big-IP, the ADC performs complex transformations within the application layer. This process can lead to problems that impact both the network and application teams, making the troubleshooting process even more complex.
• Insufficient visibility for effective load distribution – Load balancers and ADCs rely on service checks to detect server health for load distribution. Service checks run periodically, so there is an inherent undersampling problem, which can result in requests directed to servers that are failing intermittently. ADCs also tout the ability to fault in additional resources to increase scalability, but they do not offer any means to detect when this adjustment is necessary.
• Insufficient visibility for capacity planning – Load balancers and ADCs generally do not provide effective metrics on operating capacity. With little visibility into how utilization fluctuates or the remaining headroom, capacity planning becomes challenging, and network teams sometimes have to resort to guess work.

With ExtraHop, you can maximize your investment in load balancing and ADC equipment and drive optimal performance for your applications. A comprehensive L2–L7 perspective provides the visibility required to get to the heart of ADC-related performance issues.

• Tune load balancers and ADCs – ExtraHop delivers a full set of sophisticated metrics that paint a complete picture of application and network health. These advanced metrics help pinpoint delays and congestion and determine optimal settings for your specific network and application mix.
• Identify the sources of performance problems – Instead of seeing your load balancer getting blamed every time, ExtraHop can arm you with data to diagnose the true root cause, across the application, network, database, and storage tiers.
• Evaluate load distribution – Through automatic Device Groups, ExtraHop enables you to easily compare load across a pool of servers to promote even distribution. Side-by-side graphs help you spot slow servers or load-balancing deficiencies in any server pool.
• Plan for future capacity needs – With automatic trending and baselining, ExtraHop aids in capacity planning that can help you plan for future investments in application delivery controllers.