Public cloud has become a key element in many enterprises' IT plans. In growing numbers, businesses are handing their apps over to public cloud providers, such as Amazon Web Services and Microsoft Azure. And while IT pros often view security as a reason to keep apps in house, performance monitoring may be another reason, as organizations struggle to maintain visibility when transactions move off-site. Public cloud performance monitoring tools, however, can help overcome these challenges.
IT performance monitoring touches upon a complex web of interactions that are often difficult to gauge. In private and public data centers, various products, such as operating systems, network routers, storage systems and servers collect performance information. While they tell IT staff how one element, such as a server, is functioning, the ultimate goal is to assess how well end-to-end connections function.
Public cloud is playing a bigger role in these connections. Spending on public cloud services will reach $204 billion in 2016, a 16.5% increase from 2015 revenue, according to Gartner. Consequently, more companies are trying to extend their performance management systems to the public cloud -- a change that complicates end-to-end performance management.
Public cloud performance monitoring challenges
No IT performance monitoring system can trace every action. Instead, vendors have different options to evaluate performance. In some cases, they pull and interpret information from the various system components, such as network or database logs. In other instances, they place software -- or agents -- in different locations on the network, such as on a router or in a server, to send out probes. Then, they watch how long it takes for those agents to collect and return that information, and transmit it to a central location, using different algorithms to determine the status of connections.
Public cloud performance monitoring is difficult for multiple reasons. Public cloud management capabilities, in general, are less robust than those for managing on-premises services, said Julie Craig, research director at market research firm Enterprise Management Associates, based in Portsmouth, N.H.
First, the cloud provider manages its own data center and decides how much information is available to users. In some cases, the data is rich, but in others, it is limited. If a cloud service is newer, there is typically less information available. Cloud performance monitoring capabilities become more robust as providers build up their basic management functions, but if an organization customizes a cloud service, the visibility into its performance is often limited.
Another factor complicating public cloud performance management is that organizations interact with two data centers: their own and the cloud provider's. These two data centers likely have distinct, complex management systems that need to be connected. Many cloud providers and system management vendors have open application programming interfaces, so connections are possible, but because vendors orchestrate their systems in different ways, the integration work is complex.
Choosing a cloud performance monitoring tool
Despite these complexities, help is coming. Traditional IT monitoring vendors started with on-premises-based systems and, through the years, developed broad system and network management tools with capabilities for configuration management, asset management and facilities management. They also delivered performance monitoring tools that worked well with private cloud. Some of these products include BMC TrueSight Operations Management, CA Application Performance Management systems, Hewlett Packard Enterprise AppPulse, IBM Tivoli Performance Viewer and Microsoft Windows Performance Monitor.
Most of these vendors have enhanced their offerings to provide visibility into public cloud platforms, like Amazon Web Services. For instance, CA's Application Performance Management system provides monitoring for both on-premises and cloud-based infrastructure through a single view and back-end architecture.
These more traditional monitoring systems have some strengths. For example, many are familiar to existing data center staff, so minimal training is required. However, the underlying application architecture for these systems can require a time-consuming development and integration process. Also, these tools are geared toward large organizations and pricing can be high.
Newer vendors, such as AppDynamics, Cedexis, Datadog, Dynatrace, IDERA, LogicMonitor, ManageEngine, SevOne, Splunk and Zenoss, have been building cloud-friendly tools, which often cost less than traditional systems. These new tools are based on modern software, and organizations can deploy them quickly. Many of them are software as a service offerings that run in the cloud, which offloads maintenance work.
These tools can offer more niche cloud performance monitoring functions; for instance, Datadog provides deep diagnostics and monitoring for Amazon Elastic Compute Cloud Container Services. But while the functionality can be deep, it isn't always wide. Visibility into a private cloud storage system, for example, is much more limited than what is found with some of the more traditional monitoring systems. Also, these vendors are relatively small and, in some cases, are venture capital-funded -- which means their long-term future is questionable.
Public cloud is playing a larger role in IT, allowing them to more quickly deploy services and reduce management tasks. However, those benefits often come at a price: reduced visibility into system performance -- a cost that some businesses may not want to pay.
Public cloud adoption increases despite concerns
Experts make their public cloud vendor predictions
Build a cloud performance strategy