Supply: Shutterstock
VAST Information, an organization specializing in high-performance information platforms for AI and large-scale information processing, has made a few vital bulletins associated to unifying real-time information processing and storage.
The corporate has added common storage capabilities to its flagship Huge DataStore providing. It claims this makes it the trade’s first “absolutely unified” information platform for synthetic intelligence workloads.
The improve goals to get rid of the necessity for a number of, separate storage methods, permitting companies to consolidate their infrastructure and simplify information administration. The corporate stated that the brand new capabilities cater to all kinds of workload, with none trade-offs in efficiency, prices, or scalability.
“Our imaginative and prescient for the VAST Information Platform was all the time to allow organizations to course of, analyze, and act on information in real-time, empowering them to scale effortlessly, cut back their infrastructure prices, and innovate quicker by unifying AI information infrastructure inside a single, highly effective platform,” stated Aaron Chaisson, Vice President, Product & Options Advertising at VAST Information. With right now’s announcement, we’re eliminating the info silos that after hindered AI and analytics initiatives, affording prospects quicker, extra correct choices and unlocking data-driven development.”
In accordance with VAST Information, the improve permits it to fulfill Garnter’s definition of multiprotocol storage platforms, that are “designed to assist a number of storage entry protocols and tackle the rising wants of companies.”
Key options of the improve embrace assist for environments resembling VMware, Hyper-V, and different hypervisors. The platform can be optimized for Kubernetes and containerized purposes. Firms typically use each virtualized and containerized methods, which may make managing totally different storage methods tough. The VAST Information Platform has simplified this by combining each kinds of methods into one unified storage system.
The brand new options additionally embrace Boot from SAN, which permits companies to deploy and handle servers with out counting on native disks. This improves redundancy, simplifies provisioning, and enhances catastrophe restoration.
The brand new block storage performance will probably be accessible subsequent month as a part of the VAST Information Platform.
Whereas VAST’s transfer towards unified storage is promising, its block storage capabilities are nonetheless evolving. For instance, the platform helps solely NVMe-over-Ethernet presently. It lacks conventional Fibre Channel and iSCSI assist. Additionally, it doesn’t provide distant direct reminiscence entry (RDMA). Nonetheless, VAST Information has road-mapped a few of these options for future releases.
Within the different main announcement, VAST Information unveiled VAST Occasion Dealer – a real-time occasion streaming engine that unifies transactional and analytical information. The platform leverages AI brokers to dynamically act on incoming information to ship real-time intelligence and automation.
Companies right now face quite a few challenges with conventional occasion streaming methods. These older architectures typically include inflexible designs and separate information silos, making it tough to course of and analyze information effectively. This will result in inefficiencies like infrastructure sprawl and gradual information processing.
VAST Information claims that the engine delivers greater than 10x the efficiency of legacy Kafka implementations. This interprets to being able to course of over 500 million messages per second and supply limitless linear scalability.
As a unified AI information platform, VAST Occasion Dealer is designed to simplify information administration by bringing transactional, analytical, AI, and real-time streaming workloads underneath a single structure. What this implies is that as an alternative of knowledge transferring between exterior servers and slowing down, it may be streamed immediately into Huge’s DataStore.
(Yurchanka Siarhei/Shutterstock)
Commenting on the debut of the VAST Occasion Dealer, Jeff Denworth, Co-Founding father of VAST Information, emphasised its transformative impression on real-time information processing. “The launch of the VAST Occasion Dealer marks a basic shift available in the market for real-time information processing,” stated Denworth.
“By merging occasion streaming, analytics, and AI right into a single platform, VAST is eradicating a long time of knowledge pipeline inefficiencies and occasion streaming complexity, empowering organizations to detect fraud in milliseconds, correlate intelligence alerts globally, act on data-driven insights immediately, and ship AI-enabled buyer experiences. That is the way forward for real-time intelligence, constructed for the AI period.”
The brand new options permit the platform to run serverless features, which helps automate information processing and supplies real-time insights. Moreover, GPU-Direct capabilities allow quick information transfers between storage and GPUs, optimizing efficiency for demanding purposes.
The VAST Occasion Dealer is scheduled to be accessible in March.