The recent trend of cloud migration of many computing services as well as products, a lot of attention has been paid to the cloud. The majority of the focus has been on what can and should be moved, why, by whom, how, where, and when. However, little attention has been paid to the challenges related with the actual end user experience posed by migration of information and applications from dedicated to shared resources. Keep in mind that the majority of cloud benefits arise from increased sharing. This is how cost savings, scalability, availability and other benefits are achieved. However, sharing creates issues of service degradation in the form of latency, jitter, packet loss, and others.
Degradation of service is a major issue for service providers who are operating under Service Level Agreements (SLAs). The market is highly competitive and performance is the best way to differentiate and position your company. Service degradation could foretell challenges visible in cloud adoption rates much later. However, examination of the challenges and appropriate measures can allow these issues to be addressed much earlier.
The first question to ask is what are the issues faced by the service providers in seeking to assure quality services. This allows us to address the root of the problem by identifying the issues, their uniqueness, and whether the current crop of solutions can help or new ones are needed.
Before we can address this question it is important that we understand that the end user of the cloud ecosystem has certain expectations that need to be met. Virtualized services must be available, reliable, responsive, easy to use, secure, retainable, and provide utility. The end user cares about these aspects more than technical issues of infrastructure. They expect the very highest quality and will base uptake and integration decisions based on this. Therefore, meeting consumer expectations on the cloud becomes a significant issue for the future of cloud computing.
From the customer experience perspective, there are number of key indicators of quality that must be met. These need to be recognized by the service providers as well as standardized to create metrics for measurement of the quality of the service being delivered. It is critical to also recognize that these vary from one application to the other. The benefit of monitoring these KQIs is that the service provider is not only able to monitor performance and remedy issues but also they are able to meet the SLAs between them and their users.
The cloud has unique challenges that impact service quality and thereby the quality of experience for the end user. Service providers must take note that cloud-based QoE is influenced by a multiplicity of complex factors ranging from memory, virtualized compute, networking resources, and storage. These impact the ability of the cloud service provider to host the execution of applications. Furthermore, QoE is also influenced by cloud based tech components that enable the application service to function.
These resource-facing capacities bring additional risks for impairment into being. For example, an application may have to grapple with inconsistencies in delivery of resources in infrastructure due to failure of Virtual Machines, stalls, resource contention, and even premature releases. The end result is degraded quality of the application service on the user end.
This is the biggest risk to service degradation in cloud ecosystems. Consider that in a non-virtualized environment, the difference between fastest and slowest query responses is mainly insignificant. However, when executing in a virtualized environment bottlenecks result in greater service latency which poorly impacts operations. The user on their end will likely lose patience with the virtualized application as this has surpassed their KQI threshold.
Cloud service providers bring together networking, software, and other as-a-service components from many vendors to create their application. This creates a problem when seeking to trace problems and determine who bears responsibility to address the same. However, this can be addressed through standardized metrics allowing for simpler and faster accountability.
Application instances being executed in a cloud environment must place reliance on critical components delivered via partnerships to deliver good quality of experience. Examples of such components are, connectivity-as-a-service, virtual machines, and technology components. However, all of these are susceptible to vulnerabilities in terms of reliability, quality, unavailability, and latency impairments.
Applications should be designed to mitigate the impact of impairments on infrastructure. Then these applications must be tested with likely impairments in place to ensure that service quality remains acceptable and consistent. Service Level Objectives are important cloud infrastructure KQIs and also for ensuring that fault isolation is expeditious. This allows for faster accountability and therefore easier resolution of infrastructure impairments.
KQIs need to be standardized to enable faster monitoring and accelerate the maturation process across the industry. Taking into account the shared nature of the resources, focus should be on ensuring that the wait is imperceptible on the user end. Metrics, visibility, and accountability are the foundations of the solution to eradicate or mitigate impairments on cloud services.