APMdigest is following up the list of 2024 Application Performance Management Predictions with predictions from industry experts about how the cloud will evolve in 2024. Part 2 covers cloud costs, the edge and more.

Start with: 2024 Cloud Predictions - Part 1

CLOUD MIGRATIONS SLOWS

Migration to the cloud will slow down as companies reverse course. During the past few years, there has been a rapid adoption of multi-cloud strategies, with organizations often using three, four and even five different cloud environments. The problem, however, is that while organizations accumulated more cloud platforms, they ignored the problems of cross-cloud security consistency, visibility and management. A recent survey commissioned by Radware suggests that now organizations are starting to reverse course. Despite the ongoing discussion about "the great cloud migration" and the abandonment of on-premises environments, approximately three quarters of organizations not only still use these environments but expect usage to increase in the next 12 months.
Eyal Arazi
Cloud Security Manager, Radware

SHIFT FROM CLOUD TO ON-PREMISE AND EDGE COMPUTING

Cloud costs are growing in line with inflation, climate-related data center outages are continuing to increase alongside increasingly severe heatwaves and storms, and cyberattacks are now often targeting the cloud. In response to these liabilities, some companies are making a return to on-premise applications, while many others will embrace edge computing because of increased resilience due to distributed infrastructure, plus the ability to protect apps directly on the edge with web application firewalls.
Rafael Umann
CEO, Azion

Cloud Users Fully Embrace the Edge

Expanding edge infrastructure capabilities will lead to a watershed moment in 2024, where we'll see many that were previously on the cloud fully move to the edge. This will largely be due to edge infrastructure now supporting a variety of databases (including key value stores), offering the processing power needed to run AI and machine learning workloads, and enabling automating through scheduling essential edge functions to be carried out without human intervention.
Alessandro Cauduro
Chief Developer Experience Officer, Azion

Resolution of Cloud-Edge Conundrum

Predictions surrounding the domicile of enterprise data have been a roller coaster ride, oscillating between the Edge and the Cloud. The Edge initially gained the upper hand due to latency and security concerns, but the Cloud reasserted itself as latency improved. Cloud costs surged, and hyperscalers harnessed the Edge's potential, causing predictions to swing back towards the Edge. We firmly believe that the Cloud-Edge battle is about to find resolution, akin to the Mobile Core Control/Data plane separation. The Cloud, as the Control Plane, will host a multitude of applications and bulk data, while the Edge transforms into the Data Plane, catering to data primed for AI and Analytics. As this equilibrium settles, hyperscalers will extend Edge capabilities that currently reside in the Cloud. The tantalizing question remains: which Edge will reign supreme?
Ana Redondo
Product Strategy Lead, Amdocs Networks

VERTICAL CLOUD

In 2024 we will see vertical clouds continue to grow, a concept where cloud infrastructures specialize in specific industries or functions. Currently, cloud services like AWS, Google, and Azure mainly operate as general-purpose platforms, while SaaS applications cater to specific business needs. The transition from horizontal to vertical clouds is anticipated, where dedicated clouds for finance or agriculture, for instance, will offer highly specialized services. While we have seen industries starting to buy into vertical clouds, the evolution will be somewhat gradual; they won't reach their prime in 2024, but they will enjoy significant growth when it comes to adoption. Google Cloud, for example, has developed cloud offerings that cater specifically to the healthcare and manufacturing industries, and I expect this trend will continue across cloud providers and additional industries.
Manoj Chaudhary
CTO and SVP, Jitterbit

Network Exposure

The tech arena is abuzz with a pivotal shift towards network exposure. Notably, we've seen bold industry moves carving a path by establishing a separate core, untethered from CSPs, and taking control of service management. But early pioneering moves won't stand alone; many others are exploring this terrain. For CSPs, this new landscape poses challenges, as it threatens to relegate them to being mere "simple transport pipes." Brace yourselves for specific offerings poised to disrupt the status quo. Collaborations are burgeoning, weaving together the forces of CSPs, hyperscale giants, and software providers. The groundwork is laid, and 2024 promises to witness the birth of genuine monetization models that will reshape the industry.
Niall Norton
GM, Amdocs Networks

CLOUD HIGH-PERFORMANCE COMPUTING

Emerging sectors of the cloud industry becoming significant in 2024: As traditional cloud capabilities mature, Cloud High-Performance Computing (HPC) will emerge in the next 12-18 months. Current HPC workloads typically utilize on-premise supercomputing infrastructure, but the cloud providers will bring to the HPC market supercomputing capabilities wrapped in cloud-native characteristics of elasticity, programmable automation, and metered usage, democratizing the most compute-intensive scientific and engineering workloads.
Dan Krantz
CIO, Keysight Technologies

DEVELOPER ACCESS TO CLOUD

In 2024, developers will take center stage. Today's cloud tooling is mostly focused on defining and managing infrastructure (and there are amazing tools out there that keep getting better and better). But as applications get more tightly coupled with the cloud, developers are more and more frustrated that they can't leverage the full power of this platform. My prediction is that more solutions in this space will focus on giving developers greater access to the cloud and blur the boundary between application and platform.
Elad Ben-Israel
CEO and Co-Founder, Wing Cloud

PROLIFERATION OF SERVERLESS COMPUTING

Looking ahead to 2024, we'll continue to see the proliferation of serverless computing. With today's enterprises rapidly adopting more compute intensive technology, such as High-Performance Computing (HPC) and AI, we'll also see serverless support an increasing number of compute intensive workloads. Consider that while these technologies can be incredibly valuable, the costs and skills associated with these solutions can make adoption very challenging. For example, HPC users tend to have domain expertise — such as EDA, simulations, financial modeling — but they don't have the skills to provision, manage and secure infrastructure, and serverless can handle this all for them. Due to tremendous competitive pressures, AI workloads need to come to the market quickly. They're also usually very expensive, so businesses need an infrastructure that enables rapid enablement and pay-per-use models, which serverless is able to achieve.
Jason McGee
IBM Fellow, CTO IBM Cloud, GM, Platform and Common Services, IBM

VALUE-CENTRIC OPERATING MODEL

A new approach that both uplifts cloud growth and sustains the business value of cloud is critical in 2024. The focus will shift towards mastering monitoring and prediction, enabling a value-centric operating model and focusing keenly on the business case for cloud.
Anant Adya
EVP, Infosys Cobalt

CLOUD COST OPTIMIZATION

Organizations will continue to closely monitor cloud spend and optimize where possible. Third-party tools will assist them with cost optimization and likely continue to grow in popularity. The focus of the optimizations will likely be on "low value" cloud services (such as general hosting, storage, for day-to-day IT operations). There are tools included with IaaS providers but more advanced third-party tools can provide better intelligence, work across cloud providers and are more objective when reigning in spend areas that tend to run out of control if not closely monitored.
Jason Bright
Product Marketing Manager, Hyland

AI/ML DISRUPT FINOPS

AI/ML will disrupt the linear relationship in the FinOps solution ecosystem. Currently, there is a direct correlation between user complexity/friction and capabilities/cost granularity in the FinOps solution ecosystem: increases in the latter makes the former that much more complex. But advances in AI/ML will continue to disrupt the status quo, lowering the barrier and time to value for users. What used to take hours of custom configuration, trial and error, will be a low friction conversation. Additionally, AI/ML will facilitate Unit Cost solution inversion such that unit cost is provided to the user, not from the user.
Kyle Campos
CTPO, CloudBolt

GOING BEYOND FINOPS TO MANAGE CLOUD WASTE

Managing cloud costs will continue to be one of the biggest challenges and a pressing initiative for most businesses over the coming year. Economic uncertainty is forcing companies to cut budget or find efficiencies for their tech stack, and cloud waste totals tens of billions of dollars globally. Cloud consumption is particularly difficult to track and manage for enterprises with large developer organizations who have infrastructure running for different purposes throughout multiple stages of the pipeline. FinOps tools alone are no longer sufficient to tame cloud waste. Organizations will need to take a proactive, operational approach to eliminate and prevent cloud waste before it even happens, and they will need to find the right tools that can do so without stifling innovation.
Lior Koriat
CEO, Quali

Balance of power between cloud providers and users equalizes

As organizations grapple with intricate financial challenges in 2024, consumers will increasingly demand flexibility, compelling cloud providers to offer more consumer-centric pricing and services. The FinOps FOCUS initiative instituted this year is just the beginning. The collective voice of the user community will continue to be a game-changer in leveling the playing field between cloud providers and consumers.
Kyle Campos
CTPO, CloudBolt