Cloud load-balancing difficulties bog down hybrid-deployed apps

Cloud load balancing frees application server instances and provide enterprise IT with better overall performance, but current conditions make it hard in a hybrid IT environment.

Cloud load balancing improves application performance and availability, with broader overarching goals than load balancing within a corporate data center.

Consider a public cloud. Providers such as Amazon, Google, Microsoft and others all provide cloud load-balancing technologies as services. For example, Amazon Web Services' (AWS's) Elastic Load Balancing distributes workloads and traffic among Elastic Compute Cloud (EC2) instances. By comparison, Google Cloud Platform load balances network traffic between VM instances in its Google Compute Engine service.

Cloud load balancing forces enterprise IT to address security and performance in the public cloud. Support for virtual local area networks can partition cloud resources to mitigate the noisy neighbor threat of shared hardware, and bolster security for traffic. Public cloud providers also frequently mate load balancing with automatic scalability to invoke or dismiss virtual machine instances in accordance with load-balancing rules. For example, Amazon Elastic Load Balancing is often used with AWS Auto-Scaling to add or remove EC2 instances from the cloud.

Not just any app load balancer

Understand the features and capabilities of different load-balancing technologies, and which method best suits your enterprise's apps before implementing a new tool.

Hybrid cloud brings more concerns because load balancing must extend from the local data center into the public cloud cohesively. Global server load balancing and a global domain name system present a unified DNS that embraces multiple locations -- primary data center, backup data centers and the cloud providers' racks -- and determines the location that is best for serving applications to end users. Hybrid cloud load balancing frees application server instances to move provider locations depending on the business' ability to serve applications to the clients, the client device types, their locations, the value of the transactions, the costs involved and many other business-driven factors.

Such lofty goals are hard to accomplish. Most cloud providers don't expose the cloud-side performance metrics needed for load-balancing decisions. Without metrics like server response times, for example, the IT team cannot implement a least-time load-balancing architecture capable of moving traffic from local data centers to cloud instances for performance reasons.

Cloud providers also lack standards for application portability, migration and integration. Until these develop, seamless and flexible load balancing between sites and multiple providers is nearly impossible.

Next Steps

Is the "single pane of glass" a reality for hybrid cloud?

Hybrid cloud benefits are hard-earned for IT pros

Should I choose a software or hardware load balancer?

Dig Deeper on Application Maintenance on Production Systems