Traditionally, object storage — which manages data as objects in a storage pool rather than in hierarchical files or blocks — was considered best for backup use cases or data archiving.

But the furious pace of artificial intelligence  (AI) and machine learning (ML) and analytics is changing all that. Object storage is becoming a primary, favored route for demanding and fast-growing workloads — and enterprises expect scalability, high performance and quick retrieval.

Dell aims to meet these growing use cases with the expansion today of its ObjectScale tools. The company is announcing the new ObjectScale XF960, an S3-compatible, all-flash appliance built on the Dell PowerEdge server.

The new platform expands Dell’s portfolio beyond existing software-defined options and the company says it supports faster innovation, boosts performance and has built-in enterprise-grade security.

“Object is evolving very fast,” Anahad Dhillon, Dell senior manager for product management, told SDxCentral. “Our customers simply need to innovate faster.”

Object storage growing faster than ever

Gartner estimates that between 2022 and 2026, large enterprises will triple their unstructured data capacity stored as file or object storage on-premises, at the edge or in the public cloud. The firm reports that its end users see their unstructured data growing more than 30% year-over-year.

Modern enterprises struggle to store and manage this unprecedented amount of data and are turning to object-based storage, which offers scalability to terabytes or petabytes beyond the capabilities of file or block-based models.

Each object is simple and self-contained and includes data, metadata and unique identifying numbers. There are no folders, directories or complex hierarchies, and devices can be aggregated into larger pools and distributed, allowing for unlimited scale.

“Distributed file systems and object storage deployments are growing faster than ever in both volume and size as the consolidated platform for unstructured data services in global data centers,” according to Gartner.

Enterprises leveraging object storage are looking for “extensible on-premises storage products that can address an increasing number of digital business use cases with lower acquisition, operational and management costs,” Gartner says. Beyond scalability, they also expect flexibility, life cycle management, storage reliability, analytics insights and security tools.

Scaling to petabytes

Dell says it can meet these requirements with the new ObjectScale XF960, which features 11.79 petabytes of capacity per cluster, increased density and sustainability and higher performance.

“The pain points that we hear are all around scalability,” said Dhillon, adding that object storage by its nature is all about scale-out.

“You start with what you really need,” he said.  Then as data grows, users can add nodes and continue scaling out.

“Once a customer gets familiar with object storage and their development team is comfortable with updating apps, it just becomes so easy to consume,” said Dhillon.

Zero trust is ‘pillar zero’

Security is a top concern for enterprises (and across all areas). ObjectScale Lock addresses this by protecting data from unauthorized access and users can set retention policies for buckets and objects.

Dell supports identity access management (IAM) because it provides “granularity” of who accesses what data under what conditions,” Dhillon explained.

He noted that “zero trust or security in general — it’s paramount. We call it pillar zero.”

The platform supports data at rest and as well as during encryption and extends to non-security features, as well. “Security is critical for any feature that goes into the product,” Dhillon said.

For example, replication for golden copies or other data must support IAM. “No matter the functionality, it has to consider security, it has to ensure that it’s not opening any additional vulnerabilities,” said Dhillon.

Supporting AI, ML, analytics requiring millisecond latency

Dhillon said he regularly asks customers about their object storage use and hears that it goes far beyond backup and archive.

“Every single customer that I’ve talked to is running at least one business critical app on the Dell object storage platform,” he said.

While backup continues to be a primary use case, “analytics, AI, ML are the next big thing,” he said, and those use cases depend on “often-millisecond” latency. Object storage allows for “vast amounts of ingest,” he said. Large datasets can be stored in a single cluster, removing silos and allowing AI apps to read data.

“Object also allows for extensive metadata capabilities,” said Dhillon. “You can add tags, add information. It allows you to have better granularity and better AI applications.”

Other use cases include video and audio streaming, surveillance and next-gen apps built on the Amazon S3 object storage service.

“The new generation of developers are learning to develop on S3,” said Dhillon. “S3 is very easy to consume, it’s cost-effective.”

Scaling to petabytes of data

ObjectScale XF960 is built on next-generation 16G servers and allows for 33% more processing cores and memory per node, while 4X back-end network connectivity supports ingest up to 5GB per second and writes of up to 4GB per second per node.

ObjectScale manages the entire stack — storage, OS, firmware, hardware — and supports optimized data management at the edge, according to Dell. Bucket-to-bucket replication enables data redundancy and fault tolerance and data sharing across regions.

Furthermore, the platform retrieves only the data needed for an application and supports Apache Hadoop open-source to read and write S3 objects. Users can also turn on event notifications that alerts them to bucket changes (such as creation or deletion).