Cloud computing has undoubtedly changed how businesses operate, from data storage to application infrastructure. The past decade has seen explosive growth in cloud usage, but organizations now have hard-to-control cloud spending and need more options to optimize budgets.
Public cloud spending is now a significant IT expense. Roughly 80% of enterprises reported exceeding US $1.2 million per year on cloud services, and 53% of small businesses even reached the same spending on public cloud services.
Why are we spending so much? Platform-as-a-Service (PaaS) solutions have become the new way to handle infrastructure and deployment for most organizations, and usage costs quickly add up.
The current state of cloud computing is a blend of minimal options, unpredictable costs, and diminishing returns. In addition, the introduction of resource-heavy machine learning models that businesses will come to rely on further complicates the issue.
Let’s explore the dominant trends in cloud computing, what businesses are doing, and why costs have grown out of control. Keep reading to learn why we believe there’s a better way to move forward.
Approximately 80% of organizations have a hybrid cloud strategy, with only 9% using a single public cloud provider. However, this is only a step in the right direction, not the ideal solution. The same survey indicates that only 7% of organizations use multiple public cloud services.
A hybrid cloud strategy means they’re using a blend of private cloud and public cloud services, but hybrid cloud strategies often result in vendor lock-in. When your public cloud is only a single cloud provider, you’re left paying the rates they dictate, with no easy way to pivot to another provider to maximize costs.
Multi-cloud strategies go beyond the traditional hybrid cloud architecture and leverage several public cloud providers. For example, if AWS charges too much for a given task, you can switch to Azure or a local cloud provider for better rates.
The future of cloud computing is multi-cloud strategies that avoid vendor lock-in and leverage local data centers' existing knowledge and resources. This approach will allow organizations to switch between various public options on demand based on needs and current rates.
A survey of enterprise users indicates that current cloud spending is 13% over budget, and organizations expect cloud spending to increase by 29% in the next 12 months. These two statistics demonstrate that even though organizations understand they’re spending more than anticipated, they still need to keep investing in the cloud environments to grow.
But what are they getting out of these new investments? The services offered by hyperscalers have been essentially the same in the past few years. Businesses rely more on public cloud services without gaining any added benefits with new investments.
The ‘pay what you use’ nature of cloud computing was one of the early appeals that attracted enterprises, but that’s become a double-edged sword. If they use more than expected, they pay more than they’ve budgeted.
Multi cloud trends can help mitigate this by using various public cloud services based on which one offers the best rates. In addition, pivoting to cloud agnosticism will help enterprises better manage and control their spending.
Organizations self-estimate that 32% of cloud spending is wasteful, an increase of 2% from last year. Wasteful spending has become increasingly critical, especially as cloud costs increase.
Imagine if 32% of your marketing budget was a complete waste. Then, you’d revisit the campaigns and fine-tune them to ensure every dollar spent returns more than a dollar to the company.
But cloud computing is harder to manage and understand - determining the actual value it’s providing to the business is like finding the needle in a haystack. Additionally, when your only option is a single provider, there isn’t much you can do to optimize your budget and reduce waste.
Cloud providers don’t often charge to transfer data to the cloud, known as ingress; if they do, it’s a low rate.
However, moving data out of the cloud environment, known as egress, can quickly add up. Egress occurs whenever your app or platform writes data out to your on-premises environments.
Egress costs have become a significant problem. NASA conducted an internal audit of its cloud usage and concluded that egress fees create a new risk that scientific data might be less available.
You might not generate and transfer as much data as NASA, but it’s concerning that the agency may have to restrict access to critical data to save money.
Businesses are already struggling to control cloud spending, but the problem might worsen as AI and ML become relevant to businesses across most industries. It’s expected that Amazon, Microsoft, and Google will continue applying their AI technology to cloud services to meet this growing demand, but how much will it cost?
Gathering datasets and training algorithms requires a massive amount of resources, and even putting pre-trained ML models to work is resource-intensive. So will the hyperscalers start offering specialized plans for AI/ML services, or will businesses be stuck paying current rates for the significant resources they’ll start to consume to leverage these technologies?
What are businesses doing to combat cloud costs? Roughly 59% of organizations' top initiative is optimizing existing cloud usage. At the same time, the second initiative is migrating more workloads to the cloud, coming in at 57%.
This data indicates how organizations acknowledge wasteful spending but still need to rely on cloud services for even more aspects of their operations. So what’s the solution to help organizations make the most of every dollar spent on cloud services? We believe it’s investing more in IaaS and moving away from PaaS.
Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS) are two similar yet uniquely separate ways to leverage cloud computing. IaaS provides the raw infrastructure to support applications and data, while PaaS adds an extra platform layer for developers to work within.
Most organizations use PaaS due to its ease of use, which creates little room for competition. It’s incredibly costly for a new business to enter the PaaS market and compete with the size and scale of PaaS solutions like Azure or AWS.
Hyperscaler dominance results in only a few options for businesses, and they might not have an option near their headquarters. Yet, a cloud provider close to the office can deliver 50% better performance and significant savings.
Organizations can start opting for a local cloud infrastructure provider instead of a PaaS hyperscalers and cut spending without sacrificing services. A local IaaS provider will have lower latency, better hardware, faster support, and a better understanding of the regional market — all of which provide real business benefits.
It’s hard to ignore the influence of hyperscalers. Amazon and Microsoft offer appealing PaaS solutions, so developers put them to work without a second glance.
Over the past few years, this has resulted in declining developer training and experience working with IaaS services. Building a scalable cloud infrastructure with IaaS can represent significant cost savings for enterprises and SaaS platforms, but since developers are increasingly only familiar with PaaS, enterprises pay the price.
So a local data center offering IaaS solutions is left struggling to convert to a PaaS solution to cater to developers’ training and experience. But that’s often not possible, as these data centers are already operating on tight margins, and deploying a PaaS offering on the scale of AWS and Azure is cost-prohibitive.
Kubernetes has launched a new way for IaaS solutions to reimagine their services and cater to PaaS-trained developers. They don’t need to replicate AWS; they can create something new that operates similarly.
But there are still challenges for an IaaS provider to build a PaaS platform with Kubernetes, including:
Fortunately, IaaS providers don’t have to jump these hurdles alone. Lyrid’s platform tackles these challenges and creates an easy experience for developers while allowing IaaS providers to enter new markets.
Cloud computing is too important, and enterprise budgets are not infinite — hyperscalers should not be the only option. It’s time for organizations and data centers to imagine and embrace a new way to work together and launch a new era that imagines what cloud computing can provide for more reasonable and flexible rates.
Data centers worldwide already have the infrastructure in place, and many offer IaaS services. Using Lyrid’s platform, data centers can benefit from our existing technology to tackle usage monitoring, SLA reporting, and payment processing.
Enterprises can tap into Lyrid to put localized IaaS providers to work using the same training necessary to work with hyperscalers.
Ready to discover how Lyrid can help create a better future for cloud computing? Reach out to us today for a demo of our multi-cloud platform to help you optimize costs and improve operations.
99 South Almaden Blvd. Suite 600
San Jose, CA
Jl. Pluit Indah 168B-G, Pluit Penjaringan,
Jakarta Utara, DKI Jakarta