With the ongoing digitalization measures, companies have already developed a good awareness of the fact that IT budget planning and cost structures change significantly with cloud environments. However, there are critical cost traps lurking that are easily overlooked in the complex project management of a cloud migration. Apptio has identified six key areas that can break the budget of a cloud migration:

1. on-premise liabilities

When tactically planning when which workloads should be migrated to the cloud, business and strategic criteria are often at the forefront.
strategic criteria are often at the forefront. All too often, existing, decentrally managed contractual obligations associated with the existing data center are overlooked. This begins with rental contracts and ongoing hardware depreciation, as well as maintenance and support contracts, and also includes expenses in the areas of asset management, administration and audits.

It is therefore important to clarify the total costs of such on-premise contracts, which run parallel to cloud operation, in advance based on the terms. This enables a realistic cost-benefit calculation with the necessary key data to plan the cost-efficient timing of the migration to the cloud and to decide in which steps or in which order the changeover should take place.

2. active multi-cloud management

Most public cloud providers can fulfill all common cloud requirements from a single source. However, the costs and services vary in detail. It also makes sense to avoid dependence on just one provider for reasons of balanced risk diversification. In addition, multi-cloud supports data security for companies by encoding data using specific techniques and distributing it across different clouds. This means that companies - whether they want to or not - are confronted with the requirements of active multi-cloud management. Those who are technically and organizationally prepared for this with appropriate solutions, experts and processes can avoid cost traps - for example through workload analyses for better forecasting, rightsizing and cost optimization, the active use of the elasticity of cloud capacities for peak loads or through monitoring that identifies unused services or anomalies in consumption.

3. billing variants

When estimating costs and planning budgets for cloud services, it is important to understand how individual cloud
providers measure the use of their resources and the criteria they use for billing. The criteria range from hourly billing for the use of individual instances to costs based on data volume and outgoing data traffic in gigabytes per month, with different commitments.

This means that companies should have as detailed an insight as possible into which cloud capacities and qualities they require over which periods of time in the various phases of cloud migration in order to commission the optimum combination of cloud services or scaling from different providers.

Implementation of cloud strategies

4 From app performance to scaling

The architecture and performance of applications, for example in terms of memory requirements or the amount of transaction data, hardly play a role with the generally generously dimensioned hardware in data centers. In cloud operation, however, this can drive up costs surprisingly significantly.

In multi-cloud environments, it is an often underestimated cost factor when applications generate data traffic across cloud boundaries through scaling. Many cloud providers charge according to inbound and outbound traffic. Here it is important that the cost structures are known and that the applications automatically scale back again after load peaks.

The same applies to cloud storage and other resources used by DevOps teams in projects, for example. When migrating to the cloud and during operation, it is therefore helpful to carry out a cost analysis right from the start in order to be able to intervene quickly and in a targeted manner.

5 From lift & shift to legacy systems

Of the many potentially expensive changes to plans for cloud migrations, there are two in particular that can be avoided in advance by careful examination: Firstly, this concerns the popular 'lift and shift' approach, in which applications and data are moved to the cloud quickly and without adjustments and even the individual cloud services have not been dimensioned for actual use. This supposedly inexpensive approach can lead to an unexpected explosion in costs: If it is only discovered in retrospect that the costs of the public cloud infrastructure are significantly more expensive than the data center costs and architecture adjustments for the cloud are ultimately required after all.

Another cost factor can arise from old systems or legacy systems that cannot be easily migrated for technical reasons.
cannot simply be migrated for technical reasons, but must continue to be kept with their data - usually for legal reasons. As a result, on-premise resources cannot be switched off as planned and continue to burden IT budgets.

6 Exit strategy for on-premise

In the transition phase, redundant capacities in the cloud and in the data center are indispensable. This phase should be associated with consistent monitoring of workloads and analysis of actual requirements in correlation with the costs and their variables at the cloud provider(s). In addition to the technical quality, this approach ensures that decisions on when which data center capacities are switched off - or not - can be made quickly and consistently. Otherwise, there is a risk that expensive redundancies will continue for an unnecessarily long time.

The author: Thomas Köppner is a Solutions Consultant at Apptio.

© Apptio

The implementation of cloud strategies is associated with a complex interplay between consumption and costs, which is characterized by different technical and economic variables that change dynamically. It therefore requires a timely link between IT and financial monitoring with processes and systems that should be implemented at the same time as a cloud migration. This may be unfamiliar to some companies at first, but it helps to realize the optimum economic added value from cloud initiatives in the long term.