Everyone loves an open bar, but few are happy to pay to host one. Cloud costs can just as quickly get out of hand if each department in your company treats it as one, ending with an awful hangover, i.e. the bill. According to Gartner, $14.1 billion dollars were wasted last year on unnecessary cloud resources. This figure is expected to rise to 21 billion by 2021. This means over a third of global cloud expenditure is unnecessary. IT service management tools can help efficiently manage resources to reduce costs and save your IT budget.
Cloud services can be consumed easily and quickly with internet access and a credit card. But without proper control, this leads to a number of risks including:
The service catalogue plays a key role in risk minimization and cost control. It maps defined guidelines and governance requirements (cloud policy) and provides optimized approval workflows for automated ordering of cloud resources at the web, application and data level. This allows you to control which user groups may request, change and stop services and to what extent. Ideally, you should set up a self-service portal that allows users to order cloud services easily and in the same way as they order traditional services via service requests, e.g. "New PC workstation" or "Installation MS Project". This is because, in case of doubt, users do not need to know whether a service is being provided using local resources or external cloud services.
Service Level Agreements (SLAs) also form the central framework in the area of cloud computing. However, major providers such as Amazon’s AWS have standardized SLAs and their own metrics, which are not negotiable. This must be considered when selecting services.
Especially in hybrid environments, you can only ensure compliance with the SLA for a business service if the dependencies on the cloud and local infrastructure components involved are documented. However, this transparency is often not provided. Furthermore, changes in the cloud environment are often not controlled by change processes. This increases the risk of service failure.
It is therefore essential that you define, check and monitor the service levels of the cloud resources. In addition, the complete infrastructure required for an application, including its dependencies, must be documented in the CMDB and controlled via change processes, regardless of whether the resources are provided by a public cloud, private cloud or dedicated hardware ("bare metal").
The costs for cloud services must be planned, budgeted, monitored and distributed to the cost centers of the service recipients just like the costs for traditional infrastructure.
One challenge is posed by the various pricing models of cloud infrastructure providers, where billing details can vary dramatically from one provider to the next. Another special feature of external clouds is that the providers invoice the total costs of all cloud services used on a monthly basis and IT must use this information to determine the costs for the individual service recipients within the company. In order to automate the allocation process as far as possible, providers frequently offer functions for tagging the resources according to a custom naming scheme and for accessing and reading the invoices via APIs (e.g. MS Azure API). This enables users to automatically allocate cloud costs.
The creation, modification and removal of cloud resources can either be done manually or automated via a web API. As a USU study shows, even today, most providers' self-service applications are still used to manually control cloud resources (e.g. Amazon AWS console). Or the management console of the virtualization environment is used to manage local virtual machines (e.g. vSphere Host Client).
You can save a lot of work by automating this using Web APIs. However, there is no standardization of APIs because every cloud provider or hypervisor vendor relies on proprietary methods. This means that you have to build and operate an individual interface for each cloud provider or local virtualization environment. These map a variety of use cases for managing virtual machines, containers, images, snapshots, networks, etc.
The same applies to cloud migration, where workloads are moved from one cloud to another, for example in the case of a failure or when switching from a test to production environment. Within a virtualization environment, virtual machines can be moved using the management console (e.g. vSphere). Moving from a private cloud to the public cloud must be done in an orderly dismantling and rebuilding process. As with the initial deployment, migration can be performed either manually or via Web API.
In general, it makes sense to use ITIL-compliant IT service management tools for cloud management processes. This allows the central CMDB to document the current status of your entire infrastructure at any time - and thus ensure secure IT operations. Efficient cost control is equally important to avoid "cloud waste" and make optimal use of cloud resources.
In principle, cloud automation processes can also be implemented with the ITSM tool, because modern applications have all the necessary tools on board to control the various virtualization environments and cloud providers via Web API. However, if heterogeneous structures are to be controlled with virtualization environments and clouds from different manufacturers, several hundred interface use cases quickly arise. Special cloud automation solutions offer an alternative for this. These provide a large number of interfaces for the established cloud providers and virtualization platforms out-of-the-box and also offer a Web API for control by the ITSM tool.