Some workloads simply won’t move to the cloud, but the debate is heating up over what percentage stays on premises. In recent weeks, we’ve heard about cloud cost paradoxes, repatriation of workloads and earnings results from companies like Dell Technologies and HPE that indicate data center gear is in demand.
This debate over whether it makes sense to reevaluate the move to public clouds reached its peak with a blog post from Andreessen Horowitz. The argument in a nutshell: Some companies can do better if they run their own infrastructure. Think Snap, which pays billions of dollars to public cloud providers but is becoming more efficient.Think Dropbox, which took its workloads in house and weaned itself off of AWS. And think about companies where infrastructure is core to what they do.
For the rest of us, the cloud vs. on-prem decision is a bit trickier. What is clear is that public cloud spending is leaving the drunken sailor phase as Patrick Moorhead, principal of Moor Insights & Strategy puts it.
Add it up and this cloud vs. on-premises debate is really about the percentage mix. Will 25% of workloads stay on-premises or will 40%? And what mix optimizes costs? According to Flexera’s 2021 State of Cloud report, cloud costs are the hardest thing to manage. For the last five years, optimizing cloud costs was the No. 1 initiative cited in Flexera’s surveys.
Repatriation no, rethinking cloud possibly
Andreesen Horowitz’s argument was that companies that don’t go all-in on cloud can boost profit margins.
There is a growing awareness of the long-term cost implications of cloud. As the cost of cloud starts to contribute significantly to the total cost of revenue (COR) or cost of goods sold (COGS), some companies have taken the dramatic step of “repatriating” the majority of workloads (as in the example of Dropbox) or in other cases adopting a hybrid approach (as with CrowdStrike and Zscaler). Those who have done this have reported significant cost savings: In 2017, Dropbox detailed in its S-1 a whopping $75M in cumulative savings over the two years prior to IPO due to their infrastructure optimization overhaul, the majority of which entailed repatriating workloads from public cloud.
Yet most companies find it hard to justify moving workloads off the cloud given the sheer magnitude of such efforts, and quite frankly the dominant, somewhat singular, industry narrative that “cloud is great”.
The number of companies that will repatriate workloads from the cloud will likely be counted on one hand. But rest assured that more thought will be given to what data and applications stay on premises vs. move to the cloud. In addition, edge computing and multicloud deployments are rewriting some of the cloud playbook too. The big question is whether multicloud can give enterprises more leverage over vendors.
Jefferies analyst Kyle McNealy has become bullish on Hewlett Packard Enterprise largely based on the idea that not every workload will go cloud. In recent years, arguments for hybrid cloud from the likes of IBM, Dell Technologies and HPE were often dismissed.
McNealy in a research report noted the following:
- Remaining on-premises applications are getting harder to lift and shift to the cloud.
- Enterprises are getting close to their long-term on-premises footprint of 25% to 40% already.
- Edge computing, low latency and big data applications require on-premises infrastructure.
McNealy’s base case for HPE is that long-term on-premises workloads are 30%. The upside case for HPE is that on-premises infrastructure is used for 40% of workloads. In the upside scenario, HPE would benefit due to “some repatriation of compute capacity.” McNealy noted that the workload gap between public cloud and on premises will narrow.
This narrowing of the cloud vs. data center gap is showing up in the financial results of Dell Technologies and HPE. Dell CFO Thomas Sweet said:
We believe demand will continue to improve as we move through the year as customers accelerate their IT investments with focus on hybrid cloud solutions.
HPE CEO Antonio Neri said:
The overall demand environment is improving, and we are seeing traction across our portfolio. We see IT demand improving. It’s a pent-up demand to modernize that infrastructure.
Not everyone is buying the hybrid approach and repatriation argument.
Cowen & Co. published the firm’s 9th annual public cloud survey in May with 654 US and more than 800 European respondents. “COVID-19 has accelerated SaaS/Cloud migrations, with much of the impact likely permanent,” said Cowen’s analyst team.
Indeed, US respondents said spending with public cloud providers will increase 39% in 2021, up from 38% growth in 2020. Europe respondents see public cloud spending increasing 32%, according to Cowen.
One interesting nugget from the Cowen report is that respondents often underestimate the percentage of workloads moving to the cloud. Respondents are predicting 30% of workloads will remain on premises in 2023 in the Cowen survey.
It remains to be seen whether this hybrid cloud and data center renaissance continues. Data from NPD shows growth in networking and storage and declines elsewhere. It’s possible that on-premise technology spending is reverting back to 2019 levels. NPD said:
According to NPD, data center ecosystem sales in Q1 2021 were virtually flat vs. 2020 and 2019. In Q1, networking and security saw the only growth up a combined 7% and representing just under 50% of total revenue for the category in the quarter. In Q1, combined data center hardware sales fell by more than 5% year-over-year driven mainly by declines in solid state drives (-15%), hyperconverged systems (-11%), and racks, mounts and chassis (-14%). Data center ecosystem software saw continued momentum with Q1 revenue increasing 5% year-over-year and up 15% vs. Q1 2019. Software growth was again driven primarily by networking (+24%) and security (+8%).
In 2021, NPD said data center spending will continue to accelerate driven in part by edge computing and analytics, machine learning and Internet of things.
ZDNET’S MONDAY MORNING OPENER
The Monday Morning Opener is our opening salvo for the week in tech. Since we run a global site, this editorial publishes on Monday at 8:00am AEST in Sydney, Australia, which is 6:00pm Eastern Time on Sunday in the US. It is written by a member of ZDNet’s global editorial board, which is comprised of our lead editors across Asia, Australia, Europe, and North America.
