In today’s rapidly evolving IT landscape, organizations face a convergence of trends that are reshaping how they think about infrastructure—especially storage. VMware Cloud Foundation (VCF) 9.0 introduces significant advancements in VMware vSAN that align with these trends and help enterprises meet modern storage demands with efficiency, performance, and resilience.
Industry Trends and Customer Challenges
Enterprises are seeing rapid change in the data center and the marketplace, leading to challenges that increasingly can’t be met with the status quo; digital transformation is becoming an imperative.
The first trend is data growth. Data has long been thought of as digital gold, as it can provide competitive advantage through insights, new goods and new services for organizations. The type of data growing in the data center is increasingly unstructured, and overall data growth remains high. The cost to store and retain this data is a top concern. IT needs storage efficiency technologies, such as deduplication and compression, to manage total cost of ownership. They also need high capacity, low latency storage that can provide both a low TCO and the performance required by the applications that access it.
How applications access the growing volume of data is changing too. Emerging storage interfaces are being rapidly adopted across organizations that seek to turn data into new revenue streams. To meet this need, enterprises are turning to object storage and high-performance file storage as primary interfaces for AI workloads. These new workloads also demand fine-grained data management capabilities, especially for AI model accuracy and latency reduction. While IT today is adopting point solutions for these distinct data types, they’d prefer a unified solution for simplicity of management and accelerated operations.
Compounding the challenge of data growth and new storage interfaces is infrastructure sprawl via widespread adoption of hybrid cloud strategies. According to a recent study conducted by Broadcom, 92% of organizations use a mix of private and public cloud environments, which means that data and workloads are now spread across on-premises data centers, edge locations, and public clouds. Infrastructure sprawl has unfortunately spawned many unique infrastructure designs using point solutions that can be challenging to manage at scale. IT needs a consistent approach to storage infrastructure and operations across all environments to reduce complexity.
Which workloads are driving this data growth and demanding these new interfaces across environments? AI! AI workloads are placing entirely new demands on storage infrastructure. Both predictive AI, which is now well established, and generative AI, which is rapidly emerging, depend on fast, reliable access to large volumes of data. While these workloads can vary in their storage demands, even retrieval augmented generation (RAG) workloads need higher performance, lower latency storage than the legacy HDD-based storage systems currently housing much of the data AI workloads rely on. In addition, AI environments require direct, high-speed access to storage from GPUs, bypassing CPU bottlenecks and ensuring data can be processed with minimal latency.
While many apps continue to run in VMs, containers adoption continues to rise, especially for AI workloads. Containers are inherently ephemeral, but the storage they leverage must be persistent and decoupled from container lifecycles. Further, as containers proliferate, managing storage must occur at a much finer granularity—often at the level of a single container volume—and must be closely integrated with the container orchestration layer for rapid development. The bottom line is, IT needs software-defined storage that can scale at the pace of containers, and they need to manage containers with the same tools and granularity as their virtual machines. At the same time, they need to provide developers rapid access to the infrastructure via tools they use today to accelerate time to market.
Finally, cyber resilience remains a top priority. Ransomware remains a serious threat, with two-thirds of organizations suffering attacks in 2023 and over 75% of those incidents involving encrypted data. Cybersecurity commonly polls as a top pain point for CIOs.
Our Storage Vision and Today’s Solutions
VMware’s vision for storage in VMware Cloud Foundation is to deliver fully integrated, consistent storage capabilities as part of a software-defined private cloud that spans on-premises, edge, public cloud, and sovereign cloud environments. Our strategy focuses on building a single, multi-purpose storage platform that supports both primary and secondary use cases across a wide variety of workloads and data types. This platform is engineered to simplify operations, enhance security, reduce costs, and accelerate innovation.
Lowering Storage Costs
We’re uniquely addressing customer concerns about cost. VMware Cloud Foundation includes 1 TiB of vSAN storage per core in its licensing model, which has helped customers achieve an average storage cost reduction of 30%, with some reporting TCO reductions exceeding 40% compared to traditional controller-based arrays. We further reduce costs by supporting certified commodity servers, with over 500 vSAN ReadyNodes across 15 OEMs, lowering the cost per terabyte by an average of 46%. Our scale-out, hyperconverged architecture eliminates the idle capacity typically found in scale-up systems, while our server-based approach significantly reduces support and management costs. The result is high-performance, resilient storage for critical workloads—without the price tag of legacy arrays.
With VCF 9.0, we’re introducing software-based global deduplication, enabling customers to reduce storage footprints even further, reducing capacity by up to 8x — especially important as data volumes grow. vSAN’s new deduplication is designed to have minimal CPU impact and can be applied to all data within a cluster. It is a post-process dedupe, so it runs in the background opportunistically when CPU needs are low. Unlike traditional storage, where deduplication is limited to the storage behind a pair of I/O controllers, vSAN’s dedup domain scales with the cluster, potentially offering better efficiency.
In the future, we’re looking for additional ways to drive down the cost per terabyte of storage through higher density storage media that can also meet stringent performance and latency requirements. This new media, coupled with our differentiated licensing model, will unlock new use cases for vSAN and value for our customers, such as secondary storage for the backup use case.
Low Latency Storage for AI Workloads
AI workloads demand more from infrastructure, and vSAN meets that need with our next-generation Express Storage Architecture (ESA). Today, vSAN ESA delivers up to 300,000 IOPS per node with consistent sub-millisecond latency, even during peak periods. It grows both performance and capacity as the cluster scales, unlike traditional storage, with a fixed amount of performance that can only scale capacity. vSAN also handles node failure more gracefully than traditional storage, with a recent test showing 60% lower latency during a failure scenario.
In this release, we’re boosting the performance of vSAN storage clusters by up to 25% with traffic separation, enabling separate networking for compute and storage resources. This feature not only frees up networking for storage but allows customers to use existing investments. Customers looking to separate networking can potentially invest in a lower bandwidth and lower cost network for compute resources, rather than procuring high-capacity networking for compute and storage combined.
In the future, we plan to extend vSAN’s low latency storage to additional secondary storage use cases as ideal storage for predictive and generative AI workloads. vSAN will have the right mix of high capacity, low latency and scale-out storage to precisely meet the needs of these workloads.
Software-Defined Storage for All Workloads
VMware vSAN delivers a flexible, software-defined, scale-out architecture that enables organizations to optimize and expand their storage footprint on demand. Customers can scale linearly or disaggregated, incrementally to start small but achieve petabyte scale. With our CSI driver, VMware vSAN can support persistent volumes for container-based workloads, with developers able to access VMware infrastructure with the tools they use today. Admins gain granular visibility and control into container volumes, with persistent volumes as first-class citizens. As vSAN is fully integrated with the CSI driver, containers can take advantage of storage-policy based management for rapid deployment at scale, keeping pace with containers.
VCF Automation (VCF-A) in VCF 9.0 is an easy way to provide volume services to multiple tenants consuming storage resources. These storage resources can be for VMs, or persistent volumes for the use with containers in the VMware Kubernetes Service (VKS).
Looking ahead, we plan to build deeper multi-tenancy and infrastructure-as-a-service capabilities, empowering developer teams with rapid access to resources while maintaining enterprise governance and control.
Consistent Storage Operational Model Across On-Premises, Edge and Public/Sovereign Clouds
In support of consistent operations, VMware Cloud Foundation provides unified, software-defined infrastructure across all VMware environments. This simplifies deployment and management across data center, edge, and public cloud environments. VMware uniquely partners with the broadest set of public cloud partners (all major hyperscalers and hundreds of regional cloud providers) to offer our customers a consistent experience across their private cloud environments, including storage management.
VCF 9.0 includes significant innovations for storage operations. Customers can now bring all the operational benefits of VCF – automated deployments, scaling and updates – for all of their infrastructure that uses vSAN. It reduces monitoring and remediation time with multi-site capabilities, giving users a single, unified console for all of their VCF storage, including vSAN, VMFS and NFS. The unified console provides health alerts across sites, as well as guided steps to remediate any issues identified by the many health alerts built into the infrastructure.
In the near future, our hyperscaler partners will add vSAN Express Storage Architecture to their VMware offerings. VMware Cloud Foundation on Amazon Web Services, Google Cloud VMware Engine and Azure VMware Solution all will implement vSAN ESA in their VCF stacks.
In support of this strategy, vVols will be deprecated beginning with VVF/VCF 9.0 and will be fully disabled in a future VVF/VCF 9.x release. vVols did not deliver this consistent operational model across on-premises, edge, and cloud providers because it was not adopted or made available by most of our public/sovereign cloud partners. Overall, vVols adoption has remained low over the lifetime of the program, becoming a niche solution for a small segment of the vSphere customer base (low single digit percentage). See this KB article for more information. We will continue to provide our customers flexibility and choice by supporting external storage options with VMFS and NFS datastore types.
Inherent Security and Cyber Resilience for Private Cloud
Security and resilience are foundational to vSAN and an absolute necessity for VCF customers. With the introduction of vSAN 8.0, we delivered a next-generation snapshot capability, and with vSAN 8 Update 3, we delivered vSAN data protection, which allows VCF administrators to easily protect and recover VMs against accidents such as errant VM deletions, as well as malicious activities like ransomware attacks. We offer these capabilities through effortless configuration of protection groups, which allow you to define what VMs should be protected, how frequent, and for how long. The platform offers immutable snapshots and supports FIPS 140-3 validated encryption for data-at-rest and in-transit.
In VCF 9.0, we introduce vSAN-to-vSAN replication built on top of vSAN data protection to easily replicate vSAN snapshots remotely to any vSAN ESA datastore, HCI or disaggregated. vSAN replication is a lower-cost, highly performant solution for asynchronous replication, and it simplifies and accelerates recovery as compared to traditional, array-based solutions. With vSAN replication, single VMs can be replicated and restored instead of LUNs with traditional arrays, cutting down the amount of storage required at the remote site as well as the amount of network traffic. Recovery, failover and failback is much simpler than with traditional arrays as well, as VMs can be recovered as needed, not batches on LUNs. To further drive platform-level value for our customers, vSAN replication integrates with VMware Live Recovery (VLR) to enable unique benefits for cyber and disaster recovery across VCF sites. In VCF 9.0, VMware Live Recovery will support cyber restore to an on-premises VCF isolated clean room. This unlocks a key cyber recovery topology for customers who need to preserve data sovereignty to abide by data privacy or locality requirements that prevent them from storing data in the public cloud. The tight integration between vSAN replication and VLR enables a much deeper history of snapshot copies that is essential for cyber recovery, where most recent restore points have data encryption and IT teams must roll back in time to find uninfected copies. In addition, both cyber recovery and fully-orchestrated DR can be managed using a single VLR appliance to simplify operations, and users can efficiently scale their recovery infrastructure using vSAN storage clusters.
Our goal is to help our customers bridge the gap between prevention, detection and recovery. To deliver on this mission, we have announced Intelligent Threat Detection, an upcoming VLR innovation that combines AI/ML-powered proactive encryption prevention with early detection that uses data and metadata-based parameters across restore points to minimize the blast radius.
Unified Storage for All Data Types
Customers have long desired a single storage platform for all of their workloads, instead of procuring point products for particular use cases. Managing separate solutions for block, file and object storage with various point products depending on the location introduces needless complexity.
vSAN has offered block storage from day one, and in 2019 we added native file services. VCF 9.0 increases the scalability of vSAN’s integrated file services, supporting up to 500 file shares per cluster.
Future enhancements will focus on adding integrated protocols designed specifically to support the high-throughput, low-latency needs of modern AI workloads.
Looking Ahead
VMware vSAN is the integrated private cloud storage that enterprises trust to support their most critical applications. It helps organizations reduce both capital and operational storage costs, while delivering the performance, scale, and resilience needed for today’s diverse workloads—from cloud-native apps to AI-powered analytics. With consistent operations across edge, core, and cloud, and with industry-leading cyber resilience capabilities, vSAN is the foundation of the modern private cloud.
Our continued investment in innovation will drive lower TCO, simplify multi-site operations, accelerate developer velocity, and enable AI at scale to meet the most pressing challenges facing enterprise IT.
To learn more about VMware vSAN and how it can help your organization modernize storage, visit our website.
For a deeper dive into vSAN in VCF 9.0 features, check out our blogs on global deduplication in vSAN Express Storage Architecture (ESA), stretched topologies using vSAN storage clusters and network traffic separation in vSAN storage clusters.
***
Ready to get hands-on with VMware Cloud Foundation 9.0? Dive into the newest features in a live environment with Hands-on Labs that cover platform fundamentals, automation workflows, operational best practices, and the latest vSphere functionality for VCF 9.0.
