Money wasted on cloud inhibiting growth

Australian companies continue to unknowingly waste money on cloud resources they don’t need. 

As the race to adopt emerging technology like artificial intelligence (AI) becomes more prevalent, organisations are losing visibility on the cost of cloud resources needed to fuel the modernisation of IT infrastructure.

In Australia alone, organisations are expected to spend more than A$23.3bn on public cloud services in 2024, an increase of 19.7% from 2023.

For accurate and reliable use of AI models, like large language learning models (LLM) similar to ChatGPT, organisations need strong data strategies. Data needs to be clean, stored in one location and secure. The best place to do this is the cloud.

Hence the continued urgency in the uptake of cloud. For businesses to have a head-start in the AI race, they need to modernise IT infrastructure. 

Australian businesses want to be early adopters of AI, with growth identified as the most important strategic priority for businesses in the region, for the second year in a row. 

Some of Australia’s largest companies, like Commonwealth Bank of Australia (CBA) and NAB are already using AI. CBA is using AI for customer engagement, employing over 2,000 machine learning models to enhance customer interactions and identify potential scams. Meanwhile, NAB are using the tool for operational efficiency, using algorithms to match customers with suitable bankers with an emphasis on personalised services.

Growth is generally positive, however as new technologies like Arm-based processors and graphics processing units (GPUS) which enable AI capabilities are expanding the breadth of services offered, cloud environments are becoming increasingly complex.

This makes for a difficult paradox, in which organisations are pulling out check books for the cloud, putting pressure on their developers to adopt or integrate cloud technology, while losing sight of the overall budget.

If IT modernisation isn’t approached with a strategic, evidence-based plan, costs can quickly be exacerbated, inflating an organisation’s “tech debt”.

Despite the immense focus on cloud adoption and growth in Australia, a recent McKinsey report found that companies with the most tech debt diverted up to 20% of budgets earmarked for new product investment into addressing challenges related to that debt.

This creates a juxtaposition, where businesses race to invest in new technology and instead of basking in the new capabilities they offer, technology teams spend their time trying to cut down exacerbated budgets – halting innovation, growth and wasting resources. 

So where are these costs coming from?

A recent report by Datadog found 83% of container costs are associated with idle resources.

This often comes from the difficulty development teams have in accurately forecasting and right-sizing each new and existing application’s resource requirements, making it difficult to allocate resources. In addition, resource needs often change based on the nature and utilisation of workloads.

Another cause for excess cloud spends is data transfer costs – more specifically, the cost of availability zone (AZ) traffic, which refers to the traffic between resources in different availability zones often required by applications to achieve scalability.  

Some 98% of organisations are affected by AZ traffic charges. These common charges are also a great opportunity to reduce cloud cost spend, by collocating related resources into a single AZ whenever availability and application requirements allow. 

These solutions are only possible when engineers have direct access to cloud cost data. Even better, if engineers have the built-in assistance of cloud cost management software, teams can have visibility on what is pushing out cloud budgets and assess opportunities to optimise.

Traditionally, business costs are the concern of finance teams, with engineers focused on the performance of applications. But with cloud spend ramping up for most businesses, it is becoming increasingly important for engineers to have visibility across an application’s performance, cost and utilisation to help the wider business make more informed decisions.

This gives engineers, and the wider business, the ability to make cost-benefit analyses of an application’s performance and use relative to its cost. Helping organisations make informed decisions on one optimisation versus another.   

Secondly, it allows organisations to keep track of how optimisations are impacting overall costs and performance once they’ve been implemented. It offers clear insights into how optimisations are directly impacting cloud costs and performance.

The savings from visualising cloud spend, resource utilisation and performance, can be significant. Even Datadog reduced its cloud spend by A$25.9m annually. 

In total, there were 15 performance optimisations that accounted for annual savings, with the lowest optimisation yielding A$119,000 in annual cost savings and the highest optimisation yielding A$6.4m.

To use the analogy of Australia’s favourite Olympic sport – if you had the option of entering a swimming race with goggles or without, it’s common sense to choose goggles as it offers visibility that enhances your performance. 

It’s the same for cloud cost optimisation. If offered the ability to have clear oversight over cloud platforms and their costs, the choice would seem like a no-brainer. Australian businesses need to enter the cloud and AI race with full visibility to utilise resources for their intended purpose and maximise productivity while reducing excess spend. 

Siti Chen is manager of enterprise sales engineering for Australia and New Zealand at Datadog, based in Sydney