Major disruptions are expected in the cloud business this year which will see hyperscalers adapting, leaving an opportunity for challengers and startups to catch up, Forrester revealed in a new report.
Here’s what we can expect in 2024:
FinOps Open Cost and Usage Specification (FOCUS), an initiative driven by the FinOps Foundation, will become the standard in 2024, Forrester predicts. This will compel cloud providers to align on cost reporting to give customers a vendor-neutral multi-cloud view of their resources.
Even Amazon Web Services (AWS), who initially held out from FOCUS, is expected to join the community in 2024, Forrester noted.
The collective voice of the user community, driven by FOCUS, will continue to be a game-changer in leveling the playing field between cloud providers and consumers, further explained Kyle Campos, chief product and technology officer (CPTO) at CloudBolt Software.
This will be even more critical, as 2024 will be the year that the cost to run an application will be the ultimate performance metric, while traditional metrics like CPU, memory, disk and network will become less important, he added.
He said, “As digitally transformed organizations continue their advance to the cloud, the cost-of-goods-sold will get linked intrinsically with operational and services measured at a finer-grained level; looking at specific tools, services, and algorithms for cost savings.”
Further, FinOps silos that drive friction and optimization pitfalls will find a breakthrough as conversations and solutions shift from “motivation” to “facilitation”, noted Campos. This means that FinOps practices will become a native part of its “golden paths,” on par with security and observability, as the trifecta of defaults in the delivery process, especially as the majority of forward-thinking IT organizations adopt Platform Engineering as a technology approach by the end of 2024, he explained.
Advances in AL/ML will continue to lower user complexity/friction and hence cost granularity in the FinOps solution ecosystem.
“What used to take hours of custom configuration, trial and error, will be a low friction conversation,” Campos said, adding, “AI/ML will facilitate Unit Cost solution inversion such that unit cost is provided to the user, not from the user.”
2. Oracle to take on the cloud monopoly
With its increased focus on performance and cost advantages, Oracle is expected to steal hyperscaler business for more than 10 major accounts bringing at least US$100 million in annual cloud spend each, Forrester said.
The increased interest in AI will also give Oracle the opportunity to offer massive compute power to help customers train and run their generative AI models on Oracle Cloud Infrastructure (OCI). The company offers a low-latency AI supercomputer with high performance and low-cost supercluster capabilities.
3. Hyperscalers to announce new regions
By the end of 2024, public cloud vendors will have announced and launched at least two geographically separated regions in every significant market, to stay ahead of sovereignty requirements, climate change challenges, and risks connected to geopolitical tensions — a feat that will total 30 new regions across the globe, Forrester claimed.
Indeed, service outages caused by factors such as the water leak in Google’s only Paris region in April 2023, a fire-related incident at OVHcloud in 2021, or a ransomware attack wiping out CloudNordic clients’ data in August 2023 are all putting pressure on hyperscalers to failover to foreign regions.
4. Prompt engineering
Hyperscalers, Forrester said, have all the building blocks for data science except for prompt engineering.
That will change in 2024 as generative AI and foundation models continue to boom.
In 2024, all hyperscalers will announce the preview or general availability of prompt engineering, Forrester predicts. For instance, Google Cloud put Grounding Prompt in the product vision of its AI platform.
That said, enterprise adoption will be limited. Most are expected to add prompt engineering talent internally, using business subject-matter experts and data scientists to accelerate model grounding and value delivery. This is due to incomplete contextual data and limited experience in natural language and prompt engineering among data scientists. As a result, cloud providers’ first-gen prompt engineering services will not suffice to address tailored fine-tuning needs.
5. Architecture reconsiderations
Forrester predicts that the first WASI offerings will emerge on major cloud platforms and quickly spread.
WASI enables a new concept: modular monoliths, which are different from just monoliths or microservices. While monoliths are tightly integrated and efficient, but hard to iterate with agile development practices, microservices break apps into smaller chunks for faster iteration but add production complexity and require more management overhead.
WASI-enabled modular monoliths, on the other hand, will allow WebAssembly apps to run server side and introduce a new way to compose components into larger integrated applications at runtime.
The full Forrester report is available for purchase here.