X
99855

Update: 2020 Trends in Applied Infrastructure & DevOps

July 2 2020
by Brian Partridge, Henry Baltazar, Carl Lehmann, Nancy Gohring, Jay Lyman, Christian Perry, Liam Rogers, Mike Fratto, Steven Hill, Tim Stammers, John Abbott


Introduction


The coronavirus outbreak has rapidly shifted global markets, including those related to enterprise and service provider technologies, since we published our 2020 Trends report on Applied Infrastructure & DevOps in December 2019. This report serves as an update to our 2020 predictions in light of market changes due to COVID-19.

The 451 Take

As we reflect on our 2020 trends through the lens of coronavirus, short-term disruptions in demand are inevitable and extremely painful for those directly exposed to industries most impacted by COVID-19 (hospitality, retail, travel). Supply chain disruptions and social distancing have delayed the ability to physically move goods and perform tech installations. That said, as we focus on core IT infrastructure evolution, on balance these trends still hold up and in some cases are accelerating. Ample nuance is required to evaluate any short-term IT impact of this black swan event, as geographic regions, government stimulus policy and vertical industries all make a difference. As we reach the end of the COVID-19 pandemic (whatever form that takes), there will be everlasting effects on IT in both priorities and operations. We believe on balance the priorities around cloud-native technologies, DevOps, automation, cloud transformation and performance monitoring will accelerate, especially where COVID-19 exposed gaps in existing digital strategy. COVID-19 is also accelerating infrastructure investment/deployment for enterprises that can use it specifically for virus mitigation. Emergent areas such as private wireless CBRS and 5G will see short-term delays due to the physical requirements of deployment and extenuating circumstances around standards development and licensing matters. The same is true for on-premises installations in high-performance computing.

451 Research's Q4 view of the market – what we said


Figure 1
2020 Trends in Applied Infrastructure & DevOps, as of December 2019 451 Research, 2019

Our updated evaluation of each trend follows.

Business and robotic process automation markets still converging (accelerating)


In late 2019, 451 Research predicted that digital automation platforms (DAPs) designed for business process automation and robotic process automation (RPA) platforms designed for workforce task automation would coexist and gradually converge. During analysis of these two markets over recent years, we concluded that enterprises will seek capabilities to address the needs of both use cases from a common platform.

In the advent of the COVID-19 pandemic we see little evidence to alter our prediction. In fact, we see reasons why the pace toward convergence will accelerate. Many things change during a crisis that require unique and rapid responses. For example, as business travel halts and workforces shift to working from home (WFH), many business processes and worker tasks must adapt. Initially, and in particular, the IT processes needed to facilitate the shift to WFH must execute at scale, especially in IT organizations that serve global operations.

Moreover, the need for rapid response often requires making changes to existing business processes, or creating new processes all together. Under these circumstances, organizations can't wait; adaptation requires speed and agility. Many automation platform vendors have developed agile no-code technology. DAP vendors can quickly design, develop and deploy automated business processes. Likewise, RPA vendors can speed the use of software robots. Both are needed to roll out business continuity plans, establish work-at-home procedures, track medical supply chains, and manage personal health and safety. Several such vendors have responded with free access to their technology and applications to serve organizations that need new ways to cope with COVID-19. These circumstances have shined a spotlight on the need for DAP and RPA platforms to interoperate and converge for more adaptive and intelligent process automation.

As enterprise DevOps implementations move beyond mostly on-premises and private cloud environments, the hyperscale public cloud providers will become a primary and priority venue for DevOps. This comes as the hyperscalers themselves are more formally and aggressively targeting enterprise DevOps teams with their infrastructure and services. However, we do expect an impact from the Coronavirus pandemic and resulting economic conditions, which will drive more use of free, open source Kubernetes rather than paid platforms and services such as those offered by the public cloud vendors.

Enterprise DevOps will move to public clouds (accelerating)


DevOps is expanding in the enterprise at the same time that more organizations are embracing use of SaaS and public cloud infrastructures to support more application development and release processes, from testing and development to production releases. We expect 2020 will bring even more enterprise DevOps into the cloud and into hybrid and multicloud deployments. Our DevOps, Workloads and Key Projects 2020 survey indicates more DevOps is poised to move off-premises over the next two years. The survey shows on-premises, non-cloud infrastructure is the second most prevalent DevOps environment today (at 20% of respondents), behind on-premises, private cloud.

However, in two years, the on-premises, non-cloud infrastructure DevOps percentage drops to only 6% of respondents, while SaaS grows from 11% currently to 16% in two years and public cloud grows from 12% currently to 14% in two years. We don't believe the pandemic and resulting economic conditions will slow this movement substantially, if at all. Even though organizations are more interested in milking everything they can from existing on-premises infrastructure, future planning and spending centers mostly on the cloud. In addition, enterprises may be even more compelled to turn to SaaS and public cloud services in place of the burdens of attracting, hiring, training and retaining DevOps talent.

Cloud-native crossover will deepen with DevOps and data (no change)


In terms of enterprise DevOps, these teams are heavy users of cloud-native software including containers, Kubernetes, serverless and service mesh. Our VotE: DevOps survey highlights growing adoption of cloud native, with 55% of enterprise DevOps teams at full or some adoption of containers and more than 42% of DevOps teams at full or some adoption of Kubernetes. Cloud-native technology and methodology are now integral to faster software releases, more efficient IT operations, readiness and innovation thanks to advantages such as portability, consistency, scalability and abstraction amid complexity.

CBRS will disrupt the private wireless status quo (mixed)


We predicted that private LTE delivered via CBRS will gradually augment and potentially displace an incumbent install base of Wi-Fi, proprietary wireless, Ethernet and industrial wired technologies in non-carpeted enterprise locations in support of IoT devices and use cases. CBRS will also support 5G in the long run, which, given the revolutionary nature of 5G, will only provide more fuel for disruption. The majority of OT stakeholders expect to use in-private networks for 5G over the next five years, which further supports the CBRS proposition. In short, CBRS-enabled devices offer a compelling value proposition when considering coverage, mobility, cell density, device and application security, and performance. Because a mix of these capabilities are often required for IoT use cases, CBRS has wide appeal to the IoT ecosystem.

COVID-19 has slowed CBRS-enabled digital transformation projects in some industries, notably retail and public venues, where social distancing orders have halted activity. In others, CBRS activity is accelerating – especially in the healthcare segment. Memorial Health System in Springfield, Illinois, used CBRS to support its temporary triage tents and COVID-19 testing facilities.

5G won't move the needle in enterprise – planning must start now (delayed)


We've never been more dependent on broadband service, which is concurrently allowing us to work, learn, shop, entertain ourselves and socialize during the COVID-19 pandemic. In this way, the increasing focus on broadband performance at the center of our work and personal lives can be viewed as a long-term driver for 5G, especially where 5G can provide a viable alternative to inferior home broadband service. A couple of years from now, 5G networks will support ultra-low latency and mission-critical communications that enable the applications and processes supporting the digital transformation of industries, some of which will see acceleration due to COVID-19. In the short term, unfortunately, 5G rollouts face unprecedented disruption across builders, buyers, standards organizations and the government entities responsible for critical support functions, such as site licensing and spectrum auctions. 3GPP recently announced a three-month delay for the completion of Releases 16 and 17, which we have previously identified as the 'showstopping' bits of 5G. This news does not surprise us, given the group's inability to conduct face-to-face meetings until at least July. This is a major disruption that affects all 5G stakeholders globally; it should be noted, however, that this should not delay Release 15 networks because the standard is already commercialized. This impacts Phase 2 and Phase 3 of 5G rollouts, which are most relevant to enterprise deployments. Those that can dedicate resources to 5G planning during the pandemic should continue to do so as any and all 5G delays will be temporary.

AI will alter the infrastructure landscape and administration teams (no change)


The COVID-19 pandemic did not derail the points made in our 2020 Preview. If anything, it is accelerating the trend we highlighted in regards to the impact of AI on the infrastructure landscape and the administration teams that manage these environments.

The current situation has put a spotlight on the need for remote management for infrastructure systems, especially with many administrators working from home or only going into the datacenter on a limited basis. This in turn has dramatically raised the importance of reliability and proactive warnings, which are two key areas in which AI/ML-enhanced tools can provide organizations with tangible benefits.

As we mentioned in the report, AI/ML-enhanced tools are a starting point for organizations progressing toward automation, which should be the aspirational goal for most organizations given that today only 4% of respondents to our Storage, Budgets and Outlook 2020 survey claim their preferred approach to IT infrastructure management is fully autonomous remediation. In that same survey, 52% of respondents claim they experienced increased reliability and consistency from automating processes.

In conversations with IT administrators, it is apparent that some still have fears in relation to the potential impact of automation on their careers. But given that workload growth continues to be an overwhelming issue, automation will become a necessary requirement that will allow these professionals to keep up with environments that are rapidly growing and evolving.

The network will get intelligent (delayed)


Enterprise reactions to COVID-19 – namely WFH initiatives – will slow, but not stop, product and service development of AI features and adoption by enterprises. Most of the ML and AI development has occurred within cloud managed Wi-Fi services, and that work will continue as enterprises return to on-premises staffing and consumer-based businesses reopen. However, many IT plans, including implementing AI-enabled products and features within enterprises, have been put on hold as more pressing matters, such as supporting the unprecedented spike in WFH staff, take priority.

WFH initiatives, staff reductions and the closures of consumer-oriented venues will have a bigger impact on vendors and service providers that have yet to launch or are early in their AI product development plans because ML and AI algorithms are data hungry. Products and services relying on data from enterprises may have to wait for people to return to work and stores to begin the data collection and analysis to continue forward. However, product development and implementations will continue as enterprises resume more normal operations.

Pushing the limits of hyperconverged infrastructure will become a sport (accelerating)


Although not necessarily unique, an aspect of HCI that has accelerated its market momentum in recent years is the ability of vendors to relatively easily adjust or expand their platforms to accommodate customer needs. For example, disaggregated HCI continues to spread across vendor portfolios to better serve customers with scaling requirements that fall outside the scope of traditional HCI. COVID-19 now presents another opportunity for HCI vendors to quickly serve customers with strategies impacted by the pandemic, and that effort falls nicely into the core strengths of the technology.

The pandemic continues to sharpen IT managers' focus on remote work, considering that global quarantines have forced many (or most) employees to work from home or other remote locations. Key to remote work strategies is virtual desktop infrastructure (VDI), which delivers a remote, virtual desktop experience from a centralized server.

Historically, VDI succeeded primarily in well-staffed, well-funded IT environments, as scaling the three-tier infrastructure required to support the platforms was a complex, time-consuming process. HCI, on the other hand, eliminated most of those challenges with its ease of scaling – to support more users, administrators need only add more HCI nodes. Further, HCI's highly automated, intuitive management makes it easier to control costs, as it does not require specialists to manage the infrastructure.

With its recent VDI track record, HCI finds itself in an advantageous position for increased market growth. COVID-19 is likely to accelerate VDI strategies that may have been stalled or sluggish prior to the pandemic, but the unavoidable remote work requirements will deliver the appropriate justification for IT managers to get those strategies back on track.

Cloud native drives unified monitoring demand (accelerating)


We don't expect COVID-19 to affect the original trend of cloud-native adoption driving a unified approach to monitoring. Since the coronavirus crisis has shut down or drastically shrunk almost all commerce that requires a physical presence, from retail stores to restaurants to sporting events, for most businesses digital properties represent the primary channel for revenue. Some businesses are just launching or ramping up apps and services in hopes of replacing a revenue stream that is no longer viable. Others already offered digital properties but are now experiencing unprecedented demand. Either way, since poor performance leads to a loss of customers and revenue, the pressure on digital properties to perform well has escalated since the coronavirus hit.

Vendors offering products along the monitoring and incident management tool chain have an opportunity to position themselves as mission-critical support to ongoing business operations. Businesses require monitoring tools to alert them when problems are occurring and assist in quickly identifying the root cause of issues. They will rely on incident management and ticketing tools, as well as adjacent collaboration tools, that ensure they can resolve performance problems before they cause a loss of customers – and revenue.

Our discussions with vendors over the weeks since the virus hit support our analysis. They tell us that while some deals take slightly longer to close, because enterprises are scrutinizing all spending, deals are closing and pipelines look healthy. While we think that the application and infrastructure performance market will weather the storm created by the coronavirus pandemic and has the potential to thrive in its wake, vendors that are weighted toward serving small and medium businesses, and those that happen to have an outsized presence in hard-hit sectors like travel, are likely to feel some losses. Where they can make up for those losses with emerging demand, we'll see these vendors bounce back.

Merging file and object storage architectures will be either a trick or a treat (accelerating)


While the COVID-19 pandemic may be slowing some of the more non-critical development initiatives, it's likely that WFH has shined a light on the challenges of managing unstructured data that's outside the traditional enterprise firewall. This is a natural fit for hybrid cloud storage, where vendors have been shifting their message from simply providing cloud NAS capabilities to providing visibility and uniform management of data that's increasingly mobile. The appearance of COVID-19 and the ongoing WFH environment only increases the need for unstructured data governance that can span a broad range of platforms regardless of physical location, and provides a strong argument for adopting a scalable and flexible approach toward unstructured data management that combines the needs of file-based applications while providing the scalability and metadata-enabled capabilities of ubiquitous and cost-effective object storage.

Unstructured data now makes up a growing majority of business data, so from a BC/DR and data protection perspective, the COVID-19 quarantine serves as a highly visible example of the unpredictability of disasters and illustrates the fact that some of the largest interruptions in services are often based on very human conditions rather than technical failings. We believe the challenges of providing data availability, security and protection for the WFH imperative has become one of the key contributors to recent cloud storage growth, and a flexible combination of public and private cloud-based storage resources has proven to be the difference-maker in reducing the impact of the quarantine on companies that had already adopted a hybrid file and object unstructured data workflow prior to COVID-19.

New workloads, cloud native will reshape hardware (delayed)


Since we predicted in late 2019 that the era of 'good enough' general purpose systems was passing, the trend toward the establishment of next-generation accelerated infrastructure has been confirmed by two big industry acquisitions. Intel bought the AI accelerator chip startup Habana Labs at the very end of 2019 for roughly $2bn. And then in March, just as the COVID-19 lockdown was starting in the US and Europe, NVIDIA agreed to buy the networking offload company Mellanox for $6.8bn.

Once the virus kicked in, supply chain issues slowed down shipments of standard components, and therefore standard servers and storage. As for high-end on-premises systems (including high-performance computing), where installation is still an important part of the shipping process, delivery, installation and turn on was stalled due to social distancing. The extent of the damage is now becoming evident as the largest system-makers report their first full quarters of operating during the coronavirus crisis. Even while demand for new infrastructure boomed in some areas, such as remote connectivity and virtual desktops for remote working, it wasn't always possible for the suppliers to deliver, though what did happen was an affirmation that cloud services (on the whole) remained resilient and easier to access and consume as an alternative to renewing on-premises systems.

Those cloud-native services increasingly require next-generation infrastructure to handle the new demands of customers. This includes specialized and optimized processors, large memory and storage requirements, and flexible networking options to handle the exploding volumes of data required by ML and the intelligent edge. Our conclusions from Q4 remain intact, with the caveat that the transition to cloud-native services and cloudlike delivery models across the entire hybrid infrastructure estate will likely speed up – at least once current uncertainties recede a little. And the quest for rapid drug discovery, utilizing complex AI modeling, simulation and machine learning is a timely reminder of just how important having a fully modernized and optimized foundation of next-generation infrastructure in place really is.