As the modern data science stack continues to evolve, the motivations that previously drove organizations to adopt Direct Attached Storage (DAS) systems such as Hadoop Distributed File System (HDFS) have evolved. Enterprises continue to embrace real-time computing tools such as Apache Spark, Trino, and Elasticsearch, shifting away from batch-oriented tools of the Hadoop era.
To that effect, storage must be now architected for delivering real-time performance for any query load, while accommodating for the continuous evolution of data science tool-sets and the inevitable data growth. But most file and object storage systems (new and legacy) fail to scale performance affordably for your current applications, and are not architected to scale beyond a few petabytes.
VAST breaks these trade-offs with its unique Disaggregated and Shared-Everything (DASE) architecture, and provides a highly scalable and affordable all-flash, file and object platform that allows you to run petabyte and exabyte scale analysis at less than half of the cost of traditional all-flash solutions, while being many times faster.
Massively-parallel architecture delivers concurrency and real-time responsiveness at scale
Get all-flash performance for all data operations and searches at 1/2 the cost of traditional flash solutions
Use your preferred tools– Spark, Vertica, Trino, ElasticSearch, and easily adopt new tools that emerge
Scale capacity to 100s of petabytes and beyond without requiring complex capacity upgrades
Access the same data via S3 & NFS, eliminating the need for specialized storage or multiple copies of data
Turnkey appliance - with no client SW dependency, managed by VAST co-pilots (at no additional cost)
Challenge
The proliferation of ad hoc and interactive query engines means customers need predictable, real time performance for supporting thousands of queries simultaneously. Query engines need to be able to randomly read across massive data sets, which imposes significant I/O requirements, making flash infrastructure critical. This combined with the surge in Python programming and GPU-accelerated computing, means customers need to evolve beyond HDFS and traditional object storage solutions. VAST Data’s disaggregated approach to data lake storage delivers a unified , flash-optimized platform that eliminates the need for multiple systems to store different types of data, resulting in a simple & scalable environment for all your data.
A New Type of Storage Architecture
VAST Data’s Universal Storage is an enterprise grade file and object storage system, that has reimagined every aspect of what has become typical in storage system design with DASE, ultimately delivering superior scalability, resilience, and Quality of Service (QoS) at a radically lower TCO for your rapidly evolving data analytics applications.
DASE allows users to scale the performance, independently from the capacity of their system, allowing you to scale to 100s of petabytes and TB/s within the same namespace.
Get real-time responsiveness for all your data. With no east-west cluster traffic, DASE enables virtually unlimited linear, predictable scale. VAST systems in production regularly exceed 1000 GB/s.
Enable access to the same data via S3 and NFS simultaneously, eliminating the need to create multiple copies of the same data. Simply write via S3 and read that same data back via NFS or vice-versa.
Pool compute servers to provide dedicated QoS for batch, interactive, and machine learning jobs on the same namespace.
VAST’s unique similarity based data reduction provides the industry’s highest level of data reduction, allowing you to get better storage efficiency
Add enclosures and compute servers of multiple generations into a single cluster and namespace, eliminating the need for forklift upgrades needed or migration of data from the old to new clusters.
Deliver a unified analytics platform powered by an all-flash data lake to enable real-time access to all your data at scale.
Accelerate time to insights with a flash-optimized VAST data lake to power Trino’s distributed SQL query engine at scale.
Deliver a unified data analytics platform to brings all the performance and functionality of a data warehouse to a flash-powered VAST data lake.
Future-proof your Splunk storage infrastructure with a flash-optimized VAST data lake and enable sub-millisecond read latency for all Splunk data at scale
Learn More