The rise of object storage parallels the explosion in data growth. Traditional forms of storage such as block storage and file storage struggled under the onslaught of unstructured data and social media outpourings. They became unwieldy and difficult to manage. Object storage was developed as a way to solve the problem of storing very large volumes of data in a way that was scalable.
An object storage architecture provides a way to store data as self-contained units known as objects. It has become the preferred platform for the big public cloud providers. Each object contains a unique identifier, data, and any required metadata.
Block Storage, on the other hand, breaks data up into blocks of data. Each block has an identifier as opposed to each item within the block. This type of storage is used by storage area networks (SANs). File storage is the other popular format. Used heavily in direct attached and network attached storage (NAS), information is stored in discrete files using a hierarchical file directory.
While block and file storage are commonly used, object storage is now seeing rapid growth in the enterprise – businesses need its power and versatility.
How To Choose an Object Storage Solution
Object storage solutions can be distinguished by the way they approach infrastructure management and data management.
Infrastructure management encompasses IT back-office concerns like storage pooling, provisioning, load balancing, cluster scaling, hardware/system status, capacity planning, and data integrity.
Thus, it is important to look for solutions with a high degree of automation. It is probably best to avoid tools that are tied to a specific hardware platform or require manual rebalancing. What is a minor manual task at millions of files and hundreds of Terabytes becomes a major and risky task with billions of files and Petabytes of data.
Data management is more user facing. It should comprise content segregation for different tenants, content delivery for distributed clients, and ease of data ingestion. These functions include ad-hoc search, metadata customization/tagging, file viewing, and sharing. They generally handle identity authentication and access authorization, change auditing, and storage consumption metering.
As the world moves to more distributed and hybrid (on-premises and in cloud) workflows, file sizes and data sets increase, and robust tenant and data management rises in urgency. Therefore, it is important to look for object storage products that have designed with these points in mind.
Top Object Storage Vendors
In this Enterprise Storage Forum list of top object storage vendors, we spotlight the following products as leaders in the market.
- DataCore Software
- FalconStor StorSafe
- NetApp StorageGRID
- Pure Storage FlashBlade
- Dell EMC ECS
- Cloudian Hyperstore
- Hitachi Content Platform
- Scality Ring
DataCore recently acquired Caringo and its Swarm object line. The key features in Swarm include continuously checking data integrity and automatically resolving issues. Swarm was developed with scale in mind and all nodes perform all functions so recovery becomes faster as the cluster grows.
Additionally, Swarm stores metadata with files and utilizes a universal namespace for resilience and scalability. Resources can be added to Swarm in 90 seconds. Nodes can be retired non-disruptively and replaced rapidly. One company is said to manage 50 Petabytes on Swarm with a single part time resource.
- Dynamic choice of replication and erasure coding, including transitioning from one to the other as an object’s relevance decreases over time.
- Differentiated features for data management help administrators and end-users organize, find and, securely share files.
- Ability to customize metadata, search for files, view files, and generate video clips (from larger video files) from a web-based UI or via an API that can be called from custom applications or scripts.
- Tenant management features can securely segregate ownership, visibility, and access for varied data consumers.
- Permissions can be revoked at any time.
FalconStor made a name for itself in data protection. The company takes advantage of object storage in its on-premise and cloud archival offerings, with secure data containers that can take advantage of the various capabilities offered by the major object storage offerings, both on-premise, and in the cloud.
- StorSafe seamlessly adds object storage to some of its data protection products and solutions
- It uses the metadata management capabilities of object storage to access the most applicable data
- It harnesses the immutable storage of WORM-compliant offerings to provide a perpetual, always available archive
- By breaking data into fragments and dispersing them throughout the cluster, availability goes up, while a data center breach resulting in a stolen machine yields no data loss- no complete dataset can be mounted.
- FalconStor has more than an exabyte of data under management for long-term archives.
NetApp StorageGRID include flexible information lifecycle management rules, asymmetrical multi-site architecture with geo-distributed EC, and purpose-built load balancer with QoS and monitoring features.
Additionally, it integrates with NetApp’s portfolio including: FabricPool for flash to object tiering for NAS/SAN workloads, SnapMirror Cloud, and ActiveIQ. Most StorageGRID deployments support multiple tenants and workloads.
- Supports hardware appliances (disk, hybrid, and all-flash) as well as VM and software only
- Global namespace with active / active replication
- SG100/1000 load balancer reduces cost and complexity for customers (3rdparty load balancers are supported)
- Low touch operations: Ansible for automation and Prometheus metrics
- Hybrid cloud workflows (SNS, replication) to AWS, S3 compatible, and Azure
- Validated solutions with industry leaders such as Commvault, Splunk Smartstore, Rubrik, and Veeam
- All flash nodes for modern high-performance object workloads with lower latency and high throughput such as analytics and AI/ML
Pure Storage FlashBlade
As well as operating purely for object storage, FlashBlade is part of a broader solution to address organizational unstructured storage needs. It offers high throughput, low latency, high availability, and is AWS S3 API compatible.
In addition, it provides resiliency, non-disruptive upgrades, erasure coding, RAID protection, replication, and read after write consistency for applications such as analytics.
- FlashBlade is a Unified Fast File and Object (UFFO) storage platform, supporting both native high-performance file (NFS, SMB based) and Object (S3 based) storage.
- It delivers multi-dimensional performance enabling a range of applications to leverage a single FlashBlade.
- A blade-based architecture provides high throughput where multiple blades (which run the storage services) serve I/O.
- The DirectFlash module moves I/O faster to and from the upper layers of the software.
- Can serve really small objects or very large objects.
- Object SafeMode where the system including buckets and objects can be locked for a specific time period for ransomware protection or write once read many types of operations.
Dell EMC ECS
The Dell EMC ECS object storage platform offers scalability, performance, resilience, and economics to support rapid data growth. Deployable as a turnkey appliance or in a software-defined model, ECS delivers S3-compatibility on a globally distributed architecture to support workloads such as cloud-native, archive, IoT, AI, and big data analytics applications at-scale.
ECS can scale deployments to exabytes with no limits on object sizes or the number of objects in a system, namespace or bucket.
- Using multi-site federation capabilities, organizations can create globally accessible content repositories.
- As a secondary storage tier, ECS frees up expensive primary storage for business-critical apps while keeping long-term data stores readily accessible versus traditional tape archives.
- One administrator can manage billons of objects, hundreds of tenants and petabytes of data
- ECS features numerous REST-based management APIs, allowing organizations to plug into existing management and directory solutions.
- Dell Technologies has the largest installed base of unstructured data storage, according to Gartner.
- ECS supportd Splunk SmartStore analytics workloads
- Built-in data protection
- The Dell EXF900, a new all-flash appliance, includes NVMe-based SSDs on Dell EMC PowerEdge servers to deliver performance improvements and high transactions per second (TPS).
Cloudian Hyperstore offers limitless, non-disruptive scalability, mixed-configuration flexibility, consolidated file and object data, geo-distribution, and is hybrid cloud and multi-cloud ready. It integrates with all major public cloud providers.
Other management features include bucket-level policy management, eliminating cluster-wide storage policy lock-in that limits flexibility. Cloudian also offers multi-tenancy, QoS and billing for shared storage or service provider deployments.
- Ransomware protection and data security, including on-prem S3 Object LockIn addition to providing secure shell, integrated firewall, RBAC/IAM access controls, AES-256 server-side encryption for data at rest and SSL for data in transit
- Cloudian HyperStore is integrates with VMware’s vSAN Data Persistence platform. This provides a single shared storage environment for both cloud-native and traditional applications, all managed in VMware Cloud Foundation with VMware Tanzu.
- For non-VMware environments, Cloudian offers Kubernetes S3 Operator, a plug-in that enable developers to provision and manage HyperStore object storage from within their container-based applications and with no gateways or translation layers.
- Cloudian’s HyperIQ is a monitoring, observability and analytics solution for managing storage and related infrastructure across on-premise and hybrid cloud environments.
Hitachi Vantara Content Platform
The Hitachi Content Platform (HCP) provides secure software-defined object storage at exabyte scale. It harnesses various standard APIs to offer multi-cloud support, policy-based governance and compliance and metadata management. Users can take advantage of a large partner ecosystem.
Content intelligence features provide discovery and fast exploration of business data and storage operations whether on premises, off premises, in the cloud, structured or unstructured.
- Can scale from 4 nodes to 80 nodes.
- Supports S2, NFS, CIFS, REST, HTTP, HTTPS, WebDAV, SMTP and NDMP.
- Provides storage for DAS, SAN, and object.
- A cost-optimized option is available for deep data storge at massive scale
- Gateway tools extend file services to the cloud
- HCP Anywhere offers file sync and share, remote file services and data protection for a secure workplace.
Scality Ring storage runs on-premises and extends into the public cloud. It integrates file and object storage for workloads focused on high-capacity unstructured data. It encompasses multi-cloud namespaces, a native Azure object storage support, and also provides bidirectional compatibility with S3.
- Runs on commodity hardware
- Provides integrated file and object storage in one solution rather than via a gateway.
- Scale-out, peer-to-peer architecture
- Geo-replication facilitates high availability in disaster recovery.