Performance analytics in cloud platforms is becoming table stakes amid intensifying price competition for increasingly commoditized services. The ability to effectively manage application performance is a key component of every infrastructure-as-a-service (IaaS) provider’s cloud monitoring strategy. As a result, cloud platform providers must continue to add functionality due to fewer opportunities to differentiate on raw compute or storage resources.
The evolution toward distributed service-oriented architectures and cloud-based services combined with an unprecedented growth in the number of data sources – both internally and from outside the organization – has significantly increased system complexity. Yet despite the wealth of data and content, most business users continue to struggle to gain access to information they need in a timely fashion.
One of the biggest inhibitors to more widespread adoption of cloud-based services and applications is user frustration due to poor application performance. IaaS providers are consistently plagued with service outages, capacity and risk issues, and compliance pressures. These outages cost organizations millions of dollars in losses, reputational damage and fines related to GRC (governance, regulatory, compliance) violations. They also directly impact revenue and return on investment (ROI) – for the cloud platform provider and its customers.
However, many of these situations have occurred despite the presence of monitoring tools. Simply using hardware availability SLAs to manage service providers isn't effective from an end-user perspective. Software may have been too narrowly focused on a single component (i.e. servers) that it lacked the broader perspective to detect the problem in a timely fashion.
Another problem may be that static and rigid threshold-based monitoring systems trigger many false positives. This causes frustrated developers and administrators to often ignore warnings or alerts due to the low success rate of these solutions. As a result, developers spend countless hours trying to determine what caused outages or poor user experience.
Native Cloud and Mobile Apps are Key Drivers
Application performance remains a principal pain point of developers. It is not uncommon for developers to use a handful of different monitoring solutions. The problem is that the utility of each diminishes due to the incompatibility and inconsistency of the data they produce. As opposed to just network, infrastructure, application or transactional monitoring tools that are widely deployed in the enterprise, performance analytics in the cloud is more complicated.
In the cloud, operating system agents are used to collect, correlate, analyze, alert and report on issues such as host contention, network or infrastructure interruptions, cloud provider throttling or resource utilization. Cloud-based tools monitor algorithms to warn developers when service performance begins to degrade. APIs allow developers to access and integrate functionality to monitor performance across multiple applications residing in the cloud.
As more services, applications and data are natively developed and created in the cloud, or migrate to IaaS providers, better operational metrics are needed to assure performance and user experience. In addition, as more of these services and applications are accessed by an increasingly remote and mobile user community, IT departments and business unit managers need greater visibility into the performance and availability of critical applications from the end-user experience perspective.
Specifically, customers need to understand what level of performance (i.e. speed and availability) are needed from their cloud-based applications in order to deliver fast, reliable and highly satisfying end-user experiences. Customers need to be able to measure the true experiences of their most important end-user segments, including those that are far away, to ensure their cloud service provider can deliver fast and reliable experiences in key regions.
These types of tools allow businesses running on IaaS platforms such as Amazon, Google or Microsoft to monitor their service delivery and to automate development operations to reduce outages. The goal is faster, more intelligent performance data to facilitate troubleshooting to identify root cause and maintain high availability.
Performance analytics is becoming a big data service for cloud platform providers. It is already being integrated into leading cloud platforms, as exemplified by Google’s acquisition of Stackdriver. Amazon has its own monitoring service, which we expect it will enhance further through internal development. Other cloud players that hope to survive will need to follow suit. Also expect IaaS providers to continue adding functionality, such as monitoring costs and external resource utilization. These types of advanced features, which are more frequently SaaS based, are becoming table stakes as companies evaluate and choose cloud platform providers.