From the Cloud First policy established in 2010 to last year’s Cloud Smart update, it’s clear the government is driving federal agencies toward cloud computing. The strategy makes sense, as cloud has the potential to save agencies time and money while enhancing innovation.
For agencies migrating, however, the decision is less clear cut, because there are several options agencies can choose for their cloud implementation based on individual agency needs.
When it comes to cloud, federal IT managers have three service models to choose from: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). The differences are based on how much of the computing infrastructure the agency chooses to keep on premises, and who manages the infrastructure and the applications running therein.
In an IaaS environment, the agency is all in; the provider delivers an entire virtual cloud computing infrastructure, from servers and operating systems to storage and networking. That said, the agency does a vast majority of the infrastructure management. In this scenario, because the agency is doing the management, it has far more control over the environment; the federal IT team will monitor servers and storage, for example, through a vendor-provided dashboard.
The next option is PaaS. Here, the cloud vendor provides a web-based framework for federal IT developers to build their own, customized applications. The components of the environment (servers, storage, etc.) are managed by either party—the agency or the vendor—but the development team manages the applications.
For the third option, SaaS, the cloud vendor provides software only—specifically, access to software on a subscription basis—rather than providing any type of hardware or additional infrastructure components. There are a range of application types agencies can subscribe to, from email to collaboration tools to customer relationship management.
Once the agency chooses a cloud option, the next question is: how do you monitor this new environment?
For all three options, an application-centric approach to monitoring is the way to go. Even so, monitoring will differ based on the environment.
- Within an IaaS environment, there’s far more to manage. The agency doesn’t “own” the infrastructure yet is still responsible for its monitoring and management. This becomes particularly important if your federal IT team is trying to troubleshoot application issues—this is your responsibility, not the cloud provider’s.
- With SaaS applications, the agency will generally not have broad visibility into the infrastructure, and SaaS providers don’t usually provide dashboards to monitor their own applications. In addition, with SaaS applications, you’re limited to what you can monitor based on the data the SaaS provider chooses to expose through its APIs.
- Within a PaaS environment, as we discussed earlier, the development team manages the applications regardless of which organization (cloud provider or agency) is managing the infrastructure components.
The best choice is to layer on a non-vendor-specific monitoring solution to gain a vendor-neutral view. A good monitoring solution will provide an integrated view across both in-house and cloud-based applications. A great monitoring tool will go even farther by monitoring the applications from a performance perspective, factoring in the end-user experience, which will be the far more important metric in the long run.
The key is to have the right tools to help federal IT pros monitor their cloud instances and applications alongside those applications that remain in-house—ideally from a single view, with a focus on performance.