The hype is over. Cloud computing is the new normal. More than 90% of organizations have adopted cloud technologies as their primary option. IT is becoming more hybrid, integrating more cloud services.
Given the enormous complexity of modern IT infrastructure, a consolidated view of all relevant IT issues is a must. Modern solutions for hybrid cloud management offer a comprehensive view that allows overarching planning, coordination, and control of IT services. But what does that mean in practice? The following use case description shows that the integrative interplay of various disciplines such as service request, monitoring, and license management creates the necessary transparency and process automation.
Overspending and other problems – getting everyday cloud usage difficulties under control
The case is a routine one. A staff member from HR, for instance, needs a server at short notice for a proof of concept (PoC). It will be used to test a new HR software module. The system must be available the next morning. To determine the correct model, the service request system queries the various service parameters in a structured manner: how long the server or cloud service is needed, how much computing power the planned applications require, etc. As part of automated approval, the budget is checked: what funds have already been spent for what services at the cost centers in question, what funds have been requested, etc. And there is another important element: testing the extent to which the data running on the requested server is sensitive enough to require hosting in Europe or in Germany. Since critical data is to be processed in this case, corporate policy requires that the data center be located in Germany. The server is ready as ordered the very next day.
Unused cloud capacity costs 30%
But what about usage? There are frequently two scenarios for sub-optimal usage. In our case, the personnel department postpones the PoC appointment at short notice because the software evaluator has fallen ill. Or the PoC takes place as planned, but the server capacity then remains unused, and nobody remembers to shut down the server. Gartner analysts have calculated an average overspending level for unused cloud services of about 30%. Holistic hybrid cloud management solutions create the transparency necessary to identify unused capacity and are able to use this enormous leverage to ensure more economy and efficiency.
Monitoring is necessary
And how do we create that transparency? This is where a monitoring and alarm systems come in. Given a defined threshold value, the automatic monitoring and notification system identifies the cloud system as “operational but idle”. Since “up and running” entails continuous payments to cloud providers such as AWS, the applicant is now notified of the situation. The notification includes a number of alternatives for improving the situation. The system generates these suggestions automatically. The options are as follows: “Start usage as planned”, “Let us shift service to a smaller/cheaper system until you need this computing capacity”, or “Let us cancel the system in order to avoid payments to AWS (and request a new system when you really need it)”. The notification is sent to the head of the cost center, who can see from cost allocation the costs for this service right away, so that he can discuss the issue with the requester. The escalation level provides for an automatic system-side computer shutdown if certain criteria are fulfilled.
License cost factor
And what do we need to be concerned with when a server is used? In our case, the user installs a database on his AWS system and sets up the PoC environment. The selected database could be a MSSQL database, and during installation, the prescribed request process (via a software shop) was neglected due to carelessness. During installation, the user chooses to use an MSSQL database provided by AWS AND looks into the “Licenses available” option. If the system is now scanned for the first time, the installation appears in the license management module, usage is invoiced as BYOL AWS, and the associated demand is part of the compliance that is created. In this case, the customer’s cost center has no more free MSSQL licenses (or the licenses are not suitable for AWS use), and the balance becomes negative by the number of licenses needed. After a defined time threshold value has elapsed, the system starts an escalation process to address this problem.
Create SLA transparency
Another general concern is compliance with service-level agreements (SLAs). Since cloud components are frequently part of a complete service package, effective end-to-end monitoring can be ensured only with an overarching hybrid monitoring system. The corresponding SLA definitions are saved to the CMDB and real-time SLA monitoring begins automatically. The system monitors both the AWS service (availability, alerts such as cloud watch alerts, and SLA) and the overall service, including the AWS component (availability, alerts for the overall service, and SLA). The service owners are provided with dashboards for displaying status and SLA KPIs for the overall service and for the individual cloud service components. If there are deviations from the defined limit values, the responsible parties are notified immediately. An ideal system also offers extensive SLA reporting with integrated process options for commenting on disruptions, for tamper-proof corrections, and for information enrichment from ticket system information and from the report itself.
Using transparency and automation to leave shadow IT behind
In the age of the cloud, the automated use of IT solutions by specialist departments is an everyday occurrence. Current studies show that the proportion of shadow IT services averages 30%. These services arise whenever those offered by the IT department do not include the function and delivery speeds the specialist departments require. The goal of IT organizations must therefore be to counteract this tendency and offer services that specialist departments cannot reject. The items described above focus on quick, flexible, simple provision and orchestration, automated monitoring and analysis, integrated cost management, resource optimization, compliance, etc. A modern hybrid cloud management system that includes the data center environment will cover these practical requirements.
Peter Stanjeck begleitete internationale Unternehmen über 10 Jahre als Berater von Accenture bei der Optimierung von IT-Betrieben und Anwendungsentwicklungen. Heute verantwortet er als Senior Vice President der USU GmbH den Geschäftsbereich Valuemation.