Block & File Storage + Object & Cloud Storage

Software-defined storage solution removes the gap between traditional storage options.



I had the opportunity to meet with Glen Olsen, Founder of Nexustorage during the 40th IT Press Tour.


Since the introduction of object and S3 cloud storage, people have attempted to use Object as File and to scale-out File as Object. Block style workloads on Object as a use case have not been achieved -- until now.


Nexfs, the inaugural product from Nexustorage, is an automated data lifecycle solution presented as an intelligent pool of integrated tiers of Block, File, and Object/Cloud storage.


There are numerous tradeoffs between Block & File Storage (BFS) and Object & Cloud Storage (OCS):

  • BFS is expensive while OCS can be cost-effective

  • BFS is limited in scalability while OCS is massively scalable

  • BFS has file protocols while OCS has no file interface

  • BFS can present an HTTP interface while OCS has an HTTP-native interface

  • BFS has limited or no data protection while OCS has built-in data protection

  • BFS data needs to be backed up while OCS data may not require backup

  • BFS supports existing applications while OCS applications must talk to HTTP

  • BFS is high performance while OCS has lower performance

  • BFS excels with structured data while OCS is less than ideal for structured data

  • BFS cannot handle unstructured data while OCS is designed for unstructured data

While many have tried, existing solutions fail to bridge the gap between traditional and more scalable storage technology. This has resulted in misaligned storage solutions, increased costs, complexity, and business risk.


File to Object/Cloud archiving is outdated in that files are moved in their entirety between local storage and cloud/object storage. This is known as file "staging" or "caching." It leaves a stub or link in place of the original file and adds latency since the entire file is being moved even if there is just a minor change to the file.


Nexfs provides a single, addressable storage pool for three tiers of data: 1) hot data in Tier-1 SSD; 2) warm data in Tier-2 SATA; and, 3) cold data in Tier-3 cloud or object. The data goes to the most cost-effective tier based on the type of data. It breaks files into chunks and only moves a chunk of the file that has changed rather than the entire file. The chunks go to the most appropriate storage tier based on the chunk access profile.


Patent-pending Nexassert removes the need for legacy data location registers like indexes, relational databases, NoSQL databases, manifests, stub files, and symlinks. Doing so reduces risk and latency.


Intelligent data placement provides several benefits for the enterprise by reducing primary storage (Tier-1 and -2) requirements and costs, network traffic to object and cloud storage, latency on read from Object and Cloud storage, and Cloud storage transactions and costs. It also eliminates the need for third-party lifecycle management or archiving software.


Nexfs is a fully compliant Cloud and Object consuming filesystem that has been validated with more than 8,800 POSIX file operations using the pjdfstest test suite.


Advanced data management is achieved with SmartProtect which automates copies of primary storage data in self-protecting Cloud and Object storage. It replicates Nexfs POSIX file index information to local block and to Cloud and Object storage. No data recovery is required for SmartProtected data after a primary storage failure thereby reducing data protection costs.


SmartTier moves data quickly from primary storage to self-protecting Cloud and Object storage. This frees primary storage and reduces storage and data protection costs. SmartTier runs on file close, on file data chunk-change, and/or periodicity for open files.


SmartClone enables file gold images to reside on Tier-3 Object/Cloud storage by creating instant child clones from gold images. Child changed chunks remain unique to the client. Unique child changes can more to Tier-3 storage. Changes can later be forgotten (resets to gold), retained for the life of the clone, or combined with remaining gold chunks to create an entirely new file.


Key Takeaways

  • Provides a single pool of storage

  • Consumes S3 native and compatible Cloud and Object storage

  • Transparently manages the data lifecycle at the sub-file level

  • Allows customers to consume more existing Cloud and Object storage

  • Allows block style works to use Object and Cloud storage

  • Reduces primary storage needs and costs

  • Can decrease Cloud storage costs

  • No data location index, stub-files, or links