Sustainability, security, blockchained AI/ML marketplaces and federated Cloud paradigms will be the order of the year.
Prediction #1: Businesses worldwide will move to net-zero
Green initiatives are now foundational for success in public and private organizations worldwide. Sustainability is driving innovation in digital infrastructure that extends from business policies and operations to end-to-end supply chains. Therefore:
- Sustainability and technology innovation will merge
From sustainability financing to renewable energy innovations, IT organizations worldwide are rethinking their operational policies, innovating their product designs and optimizing their supply chain partnerships with business, technology and strategies that align with the planet’s climate goals.
Industry leaders will be more aggressive about integrating sustainability innovations into data center and digital infrastructure deployments and product/service development and delivery. Companies will prioritize using eco-friendly building and product designs and materials throughout their supply chains. C-level executives will make it a priority to continuously assess the biodiversity implications of business decisions for water conservation and waste reduction. - Within the decade, data centers will be powered by 100% renewable energy
An IDG survey of 2,000 global IT leaders reported that 90% of respondents identified sustainability as a priority and/or a performance metric for their organization. To ramp down the average data center power unit efficiency (PUE) from 1.59 to near 1, businesses are innovating with liquid cooling, modular power bricks, energy efficiency retrofits, and machine learning models that predict power and space usage patterns.
Advances in prime clean and renewable energy sources/stores such as fuel cells and green hydrogen are further accelerating the path to grid positivity, as is recycling data center waste heat for heating residential areas.
Global market forces are already looking beyond climate neutrality and aiming for true net-zero—no carbon at all—in their operations and across their supply chains.
This means, over the next decade, companies, governments and their partners all need to align to achieve science-based targets that will reduce greenhouse gas emissions by approximately 50% by 2030. To participate in the emerging circular energy economy, enterprises must proactively forge partnerships with energy/utility leaders and deploy more sustainable business models.
Over the next decade, all data centers will be 100% renewable. To reach net-zero, companies will pursue sustainability-focused operational excellence, innovation and reporting initiatives, which include using green power sources versus buying offsets for renewable energy and returning it to the grid. It also means reorganizing complex supply chains to prioritize vendors, products and services that enable the transition to a low-carbon world.
To enable the frictionless exchange of data and operations across ecosystems from edge to core, digital infrastructure has been predicted to be the underlying platform for all IT and business automation initiatives everywhere.
In this new world, security and sustainability will no longer be afterthoughts, as enterprises will assume that digital infrastructure services are secure and sustainable by design, and configurable through software.
Given this dynamic digital transformation landscape, here are 2022 predictions from four of our leaders:
Prediction #2: Overcoming hybrid multi-cloud complexity will dictate digital-first success
By 2023, 40% of the Forbes Global 2000 are expected to reset their cloud selection processes to focus on business outcomes rather than IT requirements.
One of the biggest challenges in this transition will be supporting the organization’s business strategy via hybrid multi-cloud, while managing increased complexity. Digital leaders who overcome cloud, data and ecosystem complexity through automation, AI/ML, APIs and edge services will gain a significant competitive edge.
- Cloud automation will accelerate digital infrastructure consumption
As the lines blur between private and public cloud and on-premises infrastructures, AI/ML-enabled cloud automation will significantly reduce management overhead and costs of public cloud infrastructure and operations. Cloud automation will also improve critical functions such as DevOps for application modernization and security for risk detection across hybrid multi-cloud architectures—accelerating labor-intensive functions for greater optimization.
Ultimately, emerging cloud automation platforms and tools will significantly advance digital infrastructure development, management, consumption and security, as everything will be software-defined. - Connected cloud ecosystems will ramp up infrastructure agility
Going forward, hybrid multi-cloud will be part of an interconnected ecosystem of diverse cloud service providers to help unlock advanced use cases and new sources of business value.
Hybrid multi-cloud environments will function as a coherent whole rather than a series of disjointed pieces, and enterprises will need to democratize cloud access and leverage AI/ML to move workloads between clouds dynamically.
Cloud-native companies will also be switching sides as they come out of the Cloud into on-premises infrastructure for greater performance and scalability, and hybrid cloud flexibilities.
To improve infrastructure agility, reduce data egress costs, and increase data protection and privacy, enterprises will move workloads out of the cloud to on-premises infrastructure and leverage Bare Metal-as-a-Service and Edge-as-a-Service offerings in a ‘cloud-adjacent’ architecture.
In a vendor-neutral ecosystem, enterprises can use public clouds as an extension of their private infrastructure and vice versa, creating the infrastructure agility to maximize the value of both.
Open source tools will continue to play a critical role in these workload migrations, while APIs will help enterprises automate the deployment of migration circuits.
Prediction #3: AI/ML at the edge will power 5G and IoT
As 5G and IoT technologies flourish, data at the edge is exploding. Information from autonomous vehicles, drones, surveillance cameras and medical IoT devices will require real-time AI/ML model inferencing at the edge. Advances will be made in the arena of legal/public policies to improve AI/ML ethics with respect to fairness, explainability and privacy protection.
- Data control and governance needs will give rise to AI marketplaces
Organizations will increasingly need to leverage external data (from public clouds, data brokers, IoT devices) to build more accurate AI/ML models. However, data providers have been hesitant to share raw data that may get used for unauthorized purposes. Similarly, data consumers are concerned about the lineage of the data and models (in transfer learning scenarios) that they are getting from external sources for security, bias, and quality reasons.
In order to maintain the chain of custody, enterprises will leverage blockchain-enabled AI marketplaces to trade data and algorithms between multiple parties in a safe and privacy-preserving manner. This will help consumers keep track of the lineage of data and AI models and provide secure enclaves at neutral locations where raw data never leaves the enclave. - Federated and wafer-scale AI will enable next-gen AI scalability
As data generated at the edge multiplies, it becomes less cost-effective and performant to move it to a centralized location for processing. For privacy and compliance control, it is also critical that data remains within an organization’s/country’s security perimeter.
Over the next five years, data gravity, latency and privacy will shift AI architectures from a centralized model to a distributed one, making distributed AI orchestrators and control planes the norm. Model training will happen in a decentralized system of distributed computational devices at the edge, with organizations shipping algorithms to the edge instead of sending raw data to a centralized location.
Models will be trained at the edge and only model weights are sent to an aggregation location to build a global AI model, thereby reducing cost and latency, and preserving data privacy.
Over the next five years, AI model training hardware solutions will become denser to handle the training needs of complex AI problems. The training hardware will consume 40KW+ of power per rack, which will require extraordinary cooling. Problem-specific ASIC/FPGA-based solutions will emerge for AI inferencing that will be more power efficient and provide better AI inference throughput.
Prediction #4: New trust models will top the cybersecurity agenda
As organizations pursue digital strategies, cybercriminals will continue their sophisticated attacks. As such:
- Converged security and zero trust will increase the difficulty and cost to attackers
Many organizations still manage security from within siloed functions and do not share the same principles and architecture necessary to help drive greater visibility and control. CISOs will prioritize breaking down silos and establishing converged cybersecurity environments.
Security will become everyone’s job; however, CISOs will lead the charge, deploying zero-trust environments that integrate IT, OT and all digital visibility and control.
Expect to see enterprises and cloud providers strengthen forces to design shared responsibility frameworks and implement solutions that deliver continuous compliance, trust, and transparency. Applying DevSecOps models and secure software development life cycles will ensure automated security throughout software development processes. - Governments will take a more active role in curbing cyber activity
Cyberattacks and ransomware demands will be considered national threats. While increased government involvement and collaboration with the private sector has the potential for good, increased regulation will make it increasingly complex for companies to operate.
Given the complexities of evolving government regulations and cybersecurity laws, enterprises will be required to include greater security expertise on their executive board. Disclosure of attacks and notice of breaches will be mandatory, such as what the US is introducing with the Cybersecurity Maturity Model Certification.
