The greater use of cloud has been built on customers selecting a single dominant public cloud provider and extending that to a hybrid on-premises deployment. This approach provides organizations with a flexible solution that meets the needs of balancing off-premises and on-premises resources yet combining them to become one effective, shared pool of resources. The other often quoted value of cloud is its ability to scale on demand and for customers to only pay for what they use. OEMs have identified the “white space” between the traditional on-premises infrastructure model and the cloud model and have entered this market with a range of on-premises IaaS offerings that compete with cloud providers’ hybrid cloud solutions.
Omdia View
The global economic uncertainty and dynamic geo-political situation has created a new dilemma for CIOs. These events have raised the age-old question of should a customer be reliant on a single vendor or approach to delivering IT services to its customers (internal and external). The answer appears to be “no” as CIOs look to spread risk by adopting multicloud and a mixture of off- and on-premises approaches. While this is not universal, it is a clear trend that Omdia’s annual Omdia IT Enterprise Insight survey (IT Enterprise Insights 2022) supports. In Omdia’s 2022 survey, 50% of respondents reported that managing security, risk, and privacy was ranked as one of the top three focus areas for 2023.
The spending intentions of companies in 2023 added further supporting evidence of this new, more diverse approach to IT delivery. In the 2022 survey, Omdia discovered 17% of the IT budget for infrastructure was allocated to physical servers and storage, and 15% to IaaS and PaaS. When this is combined with our finding that, on average, customers use three different IaaS cloud service providers and seven different PaaS cloud service providers, it is clear CIOs are spreading risk across clouds and between off- and on-premises. Therefore, this new on-premises IaaS category clearly resonates and provides CIOs with more options to spread risk in a controllable way to reduce the complexity and cost of having too many suppliers.
Key messages
- New business models have been developed by traditional hardware vendors in response to cloud computing.
- The market may be emerging, but it is gaining traction fast.
Recommendations
- Balancing the risk of using a single supplier to multiple suppliers. One of the issues with spreading risk by adopting multiple suppliers—aka not having all your eggs in one basket—is the complexity of dealing with multiple suppliers and the associated cost. Adopting an on-premises IaaS approach enables risk to be spread between where (off- and on-premises) and simplifies the who (how many suppliers are used) by offering the ability to use both on-premises and off-premises and doing this with two different cloud providers.
- Managing the skills and risk of employee churn. Another key issue for CIOs is how to manage any strategy that is overly complex in terms of the skills needed to manage the number and variety of suppliers and technologies used. By using managed services for part of the technology stack that is not directly adding business value is one approach. However, the challenge is to ensure that the use of these managed services does not become a single-point-of-failure risk in the future.
Technology overview
New business models have been developed by traditional hardware vendors in response to cloud computing
The COVID-19 pandemic has demonstrated that organizations must be agile and capable of responding to rapidly and frequently changing market conditions. Many organizations at the beginning of 2019 were planning to adopt cloud computing as a platform that could enable greater agility and flexibility, and the sudden and rapid impact of the pandemic accelerated these organizational plans. While organizations could quickly move to a cloud platform, the legacy of this sudden shift in strategy has left unforeseen consequences that organizations will be addressing over the next few years.
The biggest benefit that public cloud computing offers is its seemingly infinite ability to scale. The calculation of the value of this capability is a function of the time taken in traditional environments to increase capacity compared to the loss of business revenue due to a lack of resources. Most organizations balance this equation by carrying excess resource capacity so that business demand can be met. However, this excess capacity is a cost of unproductive resources. Figure 1 shows the on-premises-to-cloud capacity balance equation and the key TCO factors when evaluating which approach to consider.
Dell APEX, HPE Greenlake, and Lenovo TruScale operate in a slightly different way to typical cloud pay-as-you-go. These solutions have a core committed resource a customer pays for, then any additional capacity is charged per use. While the distinction is subtle, it can impact calculations of TCO if the workload is variable in its resource requirements.
The factors to consider when evaluating where a workload should be executed is one of the most complex to perform. For example, the following are just some of the key points to consider.
- Organizations need to predict the volatility (elasticity) of demand the workload will generate on the compute, memory, storage, and network resources.
- Regulatory and sovereignty restrictions on where the data must reside and the impact of this on the operational performance of the workload.
- Calculate a true total cost for the workload based on all the activities associated with developing, deploying, and managing the workload. This must take into account infrastructure, power, cooling, building, etc.
- Evaluate the metrics for a workload in terms of performance, availability, etc., and establish the required contingency plans to ensure these metrics can be successfully delivered. For example, does the workload require a DR capability, or is a backup capability sufficient?
Figure 1: On-premises IaaS model
Hardware vendors now realize that the elasticity and agility of cloud is a big driver for more adoption. Therefore, these vendors are now looking to offer more flexibility to customers. Customers have always been able to swap capital expense for revenue expense when buying or leasing infrastructure, but the new as-a-service offerings now address the elasticity requirements by providing on-premises buffer capacity that is only paid for when used.
The market may be emerging, but it is gaining traction fast
In this year’s Omdia IT Enterprise Insight survey, we asked some questions about the use of on-premises IaaS and, in particular, which approach customers were using today and what their plans were for 2024. Figure 2 shows the top-level results of this survey, where AWS Outposts is the clear early market leader, increasing its share of the on-premises IaaS market from 26% in 2022 to almost 29% by 2024.
Omdia considers the OEM vendors were late to market to compete against AWS Outposts and Microsoft Azure Stack, the two leading cloud service provider hybrid cloud offerings. The two cloud provider approaches were vastly different: AWS chose to build its own solution and sell it as an extension of its cloud services, while Microsoft elected to partner with the OEMs where its software stack was built-in to solutions from the OEMs.
What is clear is how the Microsoft approach failed to resonate with customers, as AWS Outposts will have double the market share of Azure Stack in 2024. To compound Microsoft’s challenges, the new OEM on-premises IaaS offerings appear more to customers’ liking as the combined share of HPE, Dell, and Lenovo will be more than 3× that of Azure Stack. Oracle is the latest cloud provider to enter this market with cloud@customer, and dedicated regional offerings, and we will include them in next year’s survey as named solutions.
Figure 2: On-premises IaaS 2022 compared to 2024 by vendor solution
Deeper analysis of these top-line figures shows some interesting variances.
- Microsoft Azure stack is increasing its share of the market in only 25% of vertical markets, declining by 75% by 2024.
- AWS, by comparison, is the diametrical opposite, increasing its share of the market in 75% of verticals while declining in only 25%.
- The OEMs combined are increasing in 94% of the different vertical markets, with only the telco vertical showing a decline for all three OEMs in 2024 compared to 2022.
- Lenovo is the best-performing OEM when viewed through this lens, increasing in 56% of the different vertical markets.
Figure 3 shows the analysis by company size, where large is greater than $5bn in revenue, SMB is less than $499m, and mid-size is $500m—$4.99bn. The first observation is that AWS Outposts is more popular with large enterprise customers than SMBs, and that OEM solutions are more popular with SMBs than large customers. Azure Stack is evenly distributed between all organizations irrespective of size. Omdia considers that the initial AWS Outpost configuration launched in 2019—based on AWS Outposts Rack—was too large for many SMB customers. This would account for the difference in adoption rates since the OEMs initial offerings seemed to recognize that gap, and provided smaller configurations. AWS has now expanded its portfolio ranging from 1U servers and ruggedized devices up to complete racks.
Figure 3: On-premises IaaS use by organizational size
Figures 4 and 5 show the analysis of on-premises IaaS by region for both 2022 and 2024. Figure 4 shows that North America currently prefers to use the cloud service providers’ solutions, with it being the most popular region for both AWS Outposts and Microsoft Azure Stack, and by contrast, OEM solutions are more popular in South America & the Caribbean (with North America the least popular region).
Figure 4: On-premises IaaS use by region, 2022
However, by 2024 (see Figure 5), the picture has changed slightly: the OEMs are still the most popular in South America & the Caribbean, but AWS Outposts is also now the most popular in that region, with Azure Stack the most popular in EMEA.
Figure 5: On-premises IaaS use by region 2024
Methodology
Omdia conducts numerous end-customer surveys every year, ranging from high-level IT strategy to specific-detailed technology-related surveys. The analysts also have detailed briefings with the leading technology vendors, and combine this information with that from the surveys to identify the key trends and challenges.
Roy Illsley is Chief Analyst, Cloud and Data Center Practice, at Omdia.
This article originally ran on Omdia, the research arm of Informa.
