Applications in the Private Cloud: A Viable Alternative
The term “cloud computing” is often used synonymously with “public cloud” to mean a shared network space occupied by multiple companies. Google, Amazon, and Microsoft all run large public clouds, which offer low initial costs for use and the ability to scale up services as necessary. But this common form is not the only type of cloud service available for businesses; those in need of greater control often choose a private model, which offers greater privacy and customization. The private cloud, however, also poses unique challenges for application deployment and monitoring.
The Cloud Behind Closed Doors
Private clouds can be either internally deployed and monitored or supplied by a reliable third party. In either case, there is only one “tenant” on private networks, and local IT can have as much control as they like over day-to-day operations. In comparison to the public cloud, which requires no oversight to run, this alternative requires a significant IT investment and comes with a higher initial cost for setup and deployment and unpredictable costs to run. Benefits, however, often outweigh these minor concerns, as IT pros can customize their private deployment to perform exactly as desired, secure it more easily against outside attacks, and scale up the cloud with no limitations.
Performance in the Cloud
When using a private cloud, many companies don’t think about application performance monitoring (APM), since everything is kept so close to home. The result can be chaos as multiple apps make calls for service, in turn impacting network performance and potentially limiting a company’s return on investment. An additional hurdle posed by any cloud is a breaking of traditional time-based performance metrics, as applications are virtualized either in whole or in part. IT admins, therefore, require a way to accurately monitor and report transaction performance (from any “Enter” key push to the return of data), resource consumption, and transaction volume. And while a rise in APM technologies speaks to this need in the industry, the lack of web-based, mission critical apps written to the JEE or .NET platforms poses a challenge for IT admins seeking a holistic view of their private cloud application structure.
It’s important to note a significant difference between private clouds and their public cousins, especially when it comes to performance. Applications running the public cloud are often stored entirely off-site and deployed only to local devices as needed, meaning company IT pros may see only the beginning and end of an app’s performance cycle. In private clouds, meanwhile, keeping resources close to home may create a false sense of performance surety. Both require similarly robust monitoring tools to ensure they live up to expectations.
Software-as-a-Service (SaaS), per-second APM tools have emerged to fill this management gap, tools that are able to build and graphically represent logical application topologies. In other words, they pull relational app data every second, diagram the interaction between both applications, and report their impact on the larger network. As a result, companies are able to see at a glance where bottlenecks are occurring, which apps are going outside the network for data, and which are seeing the most use. This permits issue resolution on the fly, instead of waiting for an hourly or daily download and making changes only to correct escalating problems, in effect doing damage control instead of preventative maintenance.
Cloud deployments depend on instant access and speedy response. Especially when using a private solution, companies expect seamless performance, and IT pros are tasked with justifying technology investments. It’s crucial, therefore, that companies view APM in the same way as they view the cloud: It must be simple, accessible, and always on.
Photo Credit: Wikimedia Commons
- Twitter: @NoLettering