Observing best-practice cloud cost optimisation strategies to rightsize and maintain a customer’s IT solution can result in significant cloud savings.
Cloud cost optimisation can help build purpose designed IT infrastructures that facilitate the smooth delivery of core business goals. Before migrating any company data to a virtual cloud environment, preliminary consideration should be given to the most suitable strategies relevant to the customer’s systems, applications, and data migration objectives:
‘Lift & shift’ involves relocating system and application data from an in-house server to a managed cloud environment (e.g., Azure, AWS). While lift and shift solutions often require little planning and may deliver expedited results, the process lacks flexibility and there could be surprise costs where overprovisioned workloads aren’t optimised.
Refactoring involves developing the internal setup of a computer system to maximise performance in a given environment without affecting functionality. Depending on the size of the project, the process can be complicated and lengthy (keep in mind that some vendors do not support PaaS). However, the monthly cost – not development cost – can represent savings.
Often, hybrid solutions that combine the planned implantation of both lift and shift with selected refactored assets can deliver the best-fit cloud cost optimisation strategy. Modifying applications and infrastructure to capitalise on cloud native features while rehosting wherever possible can deliver maximum cloud cost optimisation benefits.
Right size planning is an essential part of cloud cost optimisation – preparing to migrate to a cloud environment that meets the customer’s current and predicted needs will minimise wasted spend. Azure Advisor, for example, can analyse a customer’s in-house setup (perhaps 8GB RAM + 4 CPU cores) and recommend increased/decreased VM sizes suited to the customer’s workload.
Reserved instances represent a discounted billing concept in which cloud users that commit to a pre-determined level of usage could benefit from significant concessions. Savings typically range from 30% – 70% of the cost that would otherwise be associated with “on demand” pricing structures. The terms of reserved instances are likely to vary from vendor to vendor.
In some instances, cloud cost optimisation does not need to be complicated or technical. Where a customer or client does not require 24/7 access to systems and applications stored in the cloud, powering off can mean reduced spend. For example, of the 744 hours in a 31-day month (containing 23 weekdays), only 276 hours occupy the 8AM – 8PM Mon-Fri zone of high demand for access to data.
The financial benefits of maintaining suitable licensing agreements can be easily overlooked. However, an audit of all licenses and agreements could reveal areas in which hybrid licensing schemes could cut costs. Also, leveraging online storage facilities – such as SharePoint or OneDrive – can alleviate demands on paid off-site storage.
Insufficient billing structures can bring hidden issues affecting cloud cost optimisation. For example, MSPs that transact in the first instance before passing on the costs to the customer could run the risk of failing to recoup all relevant overheads in the final bill. A complete spend vs. billing audit should be completed on a quarterly basis to ensure customer billing is thorough and fair.
A typical customer’s bandwidth costs will equal around 20% of the total compute spend. Options such as ExpressRoute (which extends the customer’s on-site network into the cloud via a private connection) and Direct Connect (which facilitates peer to peer file sharing across a central hub), can help to move traffic in and out of the cloud while minimising bandwidth charges.
When an MSP acquires a new customer and begins to plan a tailored cloud cost optimisation strategy, primary considerations should revolve around whether any digital assets could be moved to lower storage tiers for instant savings (e.g., discs attached to servers must be located in the appropriate tier). Where the customer has bought into excessive high-cost data storage, there may be ‘quick wins’ via appropriate tiering.
Average costs for a single unused VM may be in the region of £100-£300 per month (forgotten S3 buckets, IP addresses, and snapshots could also bring monthly charges). Quarterly audits that result in locating, assessing, and removing any unused assets from the cloud should form a regular part of a strong cloud cost optimisation strategy.
If you require assistance with the design and roll out of bespoke cloud cost optimisation strategies for your customers, contact Synextra today.
Or visit the Synextra knowledge base for more information.
