Is the fall in the pound about to push up your AWS & Azure costs?

With the public cloud providers AWS and Azure you can choose to consume them in a local currency for example GBP £. When you do this there is a conversion from the underlying US $ rate for the service that you are consuming in the region you have provisioned it in. For example, Azure publishes their list of FX change rates here. This means you can then compare the difference between the interbank rate and the AWS / Azure native conversion rate. Then you can use this information to compare this against a credit facility like AMEX or similar.

If you have the facility in the business and you can see a difference you can translate that into a real saving and also get a minimum of 30-day terms on your cloud billing. From comparing the AWS & Azure native exchange rates with the rates available from credit providers we have seen the rates from credit providers are often better. This is a tweak and it will add up over time, but it's important to check your specific settings and circumstances for your cloud provider.

Depending on if you consume your cloud services in a foreign currency or not the FX commission charges could be a requirement or not if you consume it in your native currency.

What will be an advantage is getting a cashback card and running all of your cloud costs through this. Currently, AMEX offers a variety of cards that have cashback or reward points. The cashback card pays 0.5% on the first £10,000 and 1% on anything over £10,000 each month.

So if your cloud spend is £20,000 per month then you can get a discount on your AWS & Azure spend:

• This would translate into a saving each month of £150 on a cloud spend of £20,000 so a saving of 0.75%. This translates into £1,800 per year and that's if your cloud cost remains static.

Of course, when you look at this within a business there are probably the services that are provisioned on a monthly basis that can be put into this facility to increase the value of the saving to the business.

Every single customer we have worked with over the last decade has had cloud resources that are redundant but are still deployed. It's just the nature of changing systems on a consumption model. This could range from a deployment created for development or a customer trial that hasn't been turned off or redundant systems that are no longer in use like storage from decommissioned VMs.

If you don't already then auditing your infrastructure on a regular basis can highlight redundant or over-scoped infrastructure. Also assigning cloud infrastructure to cost centres within your business can help show where your cloud spend is going and if a particular team/customer is an unusually high proportion of the monthly cloud bill.

You can then look at ways to remove or scale things back and often save a significant amount. But this will allow you to target your efforts. This is why when we built CloudOps we included cost analysis and recommendations which are generated by the constant monitoring of your environment. We use machine learning and data from hundreds of cloud environments to build the recommendations and keep your spend optimised and as efficient as possible.

We have talked more about how to identify cloud sprawl and if you are suffering from it in this article.

The cloud providers sell the idea of a consumption-based model for their cloud services and make a big song and dance about the fact that you only pay for what you use. But how many deployments just run 24/7 all year at their full cost?

This is an area that can slash your monthly cloud spend, even if you have some auto-scaling implemented it can always be fine-tuned. For example, an organisation that works 9 to 5 systems can be scaled back two-thirds of the day and all weekend by default.

Practically if you don't have any level of auto-scaling implemented currently, then you can put processes in place for development/operations teams to turn down / off applications at certain times. Once this is in place and proven you can then use CloudOps cloud management software to automate this.

All of these are short-term high-impact steps that WILL drive your AWS & Azure cost down now. However they are not all you can do, as the founder of IG I still get my team to audit our cloud infrastructure for cost and security on a quarterly basis using our cloud management software. In this way, I know that the cloud infrastructure I am paying for is tuned to be as efficient as possible.

The CloudOps platform also has risk/licence audit lists and cost audits on a quarterly basis will identify things like new services or configuration changes such as switching to new VM types that would drive even more cost savings.

Overall I would recommend some initial steps such as the ones above, but then for the long term, implementing cloud management software would be the step I would take. This takes you from having a smart meter which shows you how much you are spending up to having a system monitoring and tuning your monthly cloud spend. CloudOps doesn't sleep and doesn't take holidays and is always watching and learning about your workloads.