Azure Storage has developed to bolster the evolution of Microsoft AI and enterprise cloud onboarding with a number of improvements.
Final January, we shared our reflection of 2023 and outlook for 2024—and what a yr it has been! Microsoft has been on the forefront of groundbreaking AI options, empowering finish clients, builders and IT Professionals within the cloud whereas accelerating enterprise migrations to Azure. Azure Storage has developed to bolster the evolution of Microsoft AI and enterprise cloud onboarding with improvements in platform infrastructure, product choices, and workload integration. A couple of notable highlights from 2024:
- Azure Storage, the filesystem for AI on Azure, enabled mannequin coaching with exabytes of information and supported inferencing/AI purposes for startups and enterprises.
- Premium SSD v2 disks, our general-purpose block storage, continued to develop double digit MoM, with expanded integration for brand spanking new workloads like Epic and Oracle.
- Azure Elastic SAN, the business’s solely cloud native SAN providing, grew by greater than 10 instances inside few months of common availability, highlighting sturdy buyer pursuits to transition from on-prem SAN to cloud.
- Azure Container Storage, our new Kubernetes storage administration answer, simplifies cloud native app improvement with a unified expertise for all backing storage choices from ephemeral disks to Azure Disks.
As we enter 2025, our focus stays on our clients, with a imaginative and prescient to offer an clever knowledge platform that maximizes the worth of their knowledge. We measure our success by our clients’ success and can proceed to develop sturdy fundamentals and improvements to learn our clients and companions. We prolong a heartfelt thanks to our clients and companions for his or her belief and collaboration, driving us to attain extra collectively.
Fundamentals for Innovation
Underneath Microsoft’s Secured Future (SFI) initiatives, we’ve strengthened safety throughout a number of classes—safety by default, design, and operations—leading to important platform enhancements in id and secret safety, community safety, engineering methods safety, monitoring, and safe design evaluation. We additionally launched customer-facing capabilities for safety and knowledge safety, together with Microsoft Defender help for Storage, Purview safety insurance policies help (in preview), Community Safety Perimeter (in preview), and TLS 1.3 Blob help. Now we have plans in 2025 to broaden our safe by default posture throughout numerous choices, with insurance policies leveraging identity-based authentication and discouraging nameless entry.
Our investments in platform resiliency deal with proactive mitigation of impactful buyer points by way of predictive evaluation. We achieved 52% discount in excessive influence incidents within the second half of 2024 in comparison with 2023. In parallel, we proceed to empower clients in constructing extremely accessible (HA) options and enhancing our enterprise continuity and catastrophe restoration (BCDR) help by way of capabilities like multi-volume snapshots for Premium SSD v2/Extremely Disks, and deliberate failover for Blobs and Recordsdata (in preview). In 2025, we are going to leverage AI and construct automation to additional improve failure prediction and detection, lowering each the prevalence and restoration time of excessive influence incidents; and empower clients with richer insights to trace workload well being indicators for proactive mitigation.
Improvements for AI and rising workloads
From powering cutting-edge AI options to driving seamless cloud experiences, storage innovation has develop into the spine of technological progress. However how do organizations keep forward on this quickly evolving panorama? The reply lies in centered improvements—focused developments that rework how we handle, entry, and make the most of knowledge.
Accelerating AI innovation with differentiated capabilities
In 2024, we superior massive language mannequin improvement in collaboration with OpenAI, with Azure Storage serving as the inspiration for the AI lifecycle. At Ignite, we showcased how OpenAI leveraged Azure Blob Storage all through the lifecycle, from ingesting huge datasets for mannequin coaching to enabling seamless knowledge administration and enlargement with out growing operational complexity. Such achievement is made doable with Blob Storage scaled accounts offering extremely scalable throughput to forestall efficiency bottlenecks and simplify infrastructure. Extra of those capabilities that we’ve developed in collaboration with AI-scale clients corresponding to OpenAI will are available in 2025. We may even convey these developments into Azure Knowledge Lake Storage (ADLS) and supply extra clever capabilities on ADLS concentrating on AI workloads for extra environment friendly knowledge processing.
We’re engaged on an thrilling innovation that may seamlessly join unstructured knowledge with Azure AI companies and different firms’ AI options, whether or not that knowledge resides on-premises, in different clouds, or inside Azure. Constructing multi-modal enterprise AI purposes requires entry to huge quantities of unstructured knowledge however enabling safe and well-governed entry to important enterprise knowledge throughout cloud and on-premises environments is a serious problem for patrons. This rising answer will sort out the problem by securely connecting your knowledge with Azure AI companies like Azure AI Foundry, and Azure Machine Studying, via their deep integration with Azure Blob—with out the necessity for knowledge migration or copying. By accelerating entry to AI, lowering time to perception, and enabling the logical unification of information from a number of sources, this answer will pave the best way for transformative developments in enterprise AI. With strong governance powered by instruments like Azure RBAC, it offers you full management over your knowledge property. The restricted preview of this answer will probably be accessible quickly, with common availability deliberate for 2025.
Centered improvements for rising workloads
Because the cloud-first method turns into the default amongst startups and enterprises, we’re witnessing the emergence of latest data-centric workloads associated to AI coaching, safety scanning constructed on stateful containers. In response to growing demand of storage catered for containers, we launched Azure Container Storage, which presents a spread of storage choices by way of the native Kubernetes interface, with help for brand spanking new block storage choices like ephemeral disks and Elastic SAN. That is particularly useful for workloads hosted on digital machines (VMs) with graphics processing items (GPUs), the place terabytes of native NVMe can speed up mannequin coaching and inferencing. Furthermore, builders now not want to decide on between resiliency and pace for ephemeral disks. Azure Container Storage presents replication with versatile choices, making certain excessive resiliency inside or throughout zones and finest efficiency in the marketplace. In 2025, we are going to improve integration with Azure Kubernetes Service (AKS) by streamlining the set up expertise and enhance cost-efficiency with common availability help for Elastic SAN.
Optimizations for mission important workloads
We proceed to optimize the expertise for operating mission-critical workloads throughout our product portfolio. We dropped at the market—Azure Elastic SAN, optimized for price-performance at scale and seamless on-premises SAN migration to Azure. With the common availability of Azure VMware Answer integration, clients can leverage it not just for VMs and containers but additionally as a local VMware datastore. We lately launched 99.99% uptime SLA help, on the spot snapshot (in preview) with Azure Backup (in preview), and capability auto-scaling (in preview). These capabilities solidify Elastic SAN as a number one alternative for enterprise storage at scale. Trying forward, we’re specializing in optimization for database and bare-metal use instances, lowering entry latency and bettering IOPS scalability, and advancing BCDR help.
In parallel, we’re advancing our Disk choices to fulfill the rising wants of enterprises. To assist clients scale up database efficiency throughout peak site visitors, we launched Mbv3 VM collection, supporting as much as 650,000 disk IOPS—greater than a 50% enhance. We additionally enabled common availability for reside resizing on our Premium SSD v2/Extremely Disk choices, simplifying on-demand storage enlargement. With the added help for disk kind conversion, clients can improve to those newest SKUs in place. Within the coming yr, we’re touchdown extra enhancements on Premium SSD v2/Extremely Disks together with automated catastrophe restoration by way of Azure Web site Restoration (ASR), and Azure Migrate integration.
Azure NetApp Recordsdata (ANF) brings the energy of NetApp ONTAP to the cloud as a local Azure service, providing broad protocol compatibility, and sub-millisecond latency. The overall availability launch of enormous volumes, elevating the size from 50 TiB to 2 PiB with over 700,000 IOPS and 12.5 GiBps throughput, positions ANF to deal with mission important workloads in sectors like Digital Design Automation and high-performance computing (HPC). We’ll proceed optimizing cost-efficiency and enabling integration with Azure AI companies.
Azure Recordsdata presents absolutely managed file shares for hybrid, lift-and-shift and cloud native purposes. To convey you the advantages of pay-as-you-go pricing whereas offering predictable prices, we launched a brand new provisioning mannequin—Provisioned v2 on Commonplace Recordsdata. You possibly can independently scale out the dimensions, IOPS, and throughput per workload necessities, ranging from 32 GiB. In 2025, our focus is to ease raise and shift of workloads with simplified administration and seamless knowledge migration whereas enabling cloud native use instances with built-in id, and enhanced efficiency.
Main with our accomplice ecosystem
We proceed to strengthen our accomplice ecosystem to offer clients with alternative and ship strong hybrid administration capabilities adhering to excessive requirements for high quality and safety. In 2024, we collaborated with companions in delivering Komprise clever tiering for Azure, Qumulo Azure Native v2 and Cloud Native (launched modern value/efficiency tiering between Premium v2 Disks and Blob), and Veeam Knowledge Cloud Vault constructed on Blob. We satisfaction ourselves on joint innovation and are excited to announce upcoming releases in 2025:
- Nasuni Knowledge Service will permit our joint clients to entry knowledge saved in Nasuni UniFS by way of a Blob API appropriate gateway.
- Pure Storage will launch a completely managed scalable block storage service on Azure.
- Dell will convey Dell-Managed as-a-Service choices of OneFS and Knowledge Area to our joint clients.
Trying forward with Azure Storage
As we step into 2025, the horizon for storage innovation is extra thrilling than ever. With developments like AI-driven optimization, sustainable infrastructure, and lightning-fast entry speeds, we’re redefining what’s doable in Azure Storage. The yr forward isn’t nearly storing extra—it’s about empowering companies to attain extra with smarter, scalable, and future-ready storage options. Right here’s to a yr of turning knowledge into potentialities!
