File-based workloads are at the heart of innovation and of collaborative workflows in many enterprises. Enterprises pursuing transformation increasingly store multiple petabytes of unstructured file data. They frequently need to process that file data in the cloud and to access and collaborate on that data from many different locations. Public cloud providers offer some file services of their own, but have welcomed other file service providers into their clouds because those providers help move more workloads into their clouds.
New Dimensions of Scalability for File Infrastructures
Enterprises capture file data from more sources and in a greater variety of file sizes than in the past. Many enterprises are now running workloads that until recently were the domain of academic and scientific high-performance computing (HPC) environments. Many of these workloads rely on file-based storage. These include video content creation and management, surveillance, AI/ML, autonomous vehicle technology, radiology, oil and gas exploration, and more.
Multiple reasons enterprises are seeking modern file infrastructures
- Workload migration to the cloud
- AI/ML (in some cases on data they used to throw away)
- Integrating new data streams into their business processes
- File sizes, file counts, and file volumes that exceed the limits of legacy infrastructure
- Distributed operations
- The need to simplify infrastructure management and drive out costs while providing services to a distributed workforce
Multiple approaches to hybrid cloud file services
- Deploy or rent capacity on traditional filers in colocation facilities near the public cloud providers to provide hybrid/multi-cloud.
- Deploy traditional filers in public cloud data centers
- Cloud-native file services running directly on public cloud infrastructure
- Edge caching in front of traditional filers or cloud-native file services
Alternatives for hybrid cloud file services
Qumulo’s software-defined hybrid file system provides high-performance file services for trillions of files in multi-petabyte data sets. This is true on-premises, natively in the public cloud, and as a system that scales across on-premises and public cloud environments. It provides the scalable data services and real-time visibility customers need to manage the ever-increasing volume, velocity and variety of data in the enterprise.
Nasuni’s cloud-native global file system, UniFS, unifies the capabilities of enterprise NAS, backup solutions, and disaster recovery infrastructure. It uses object storage as the backing store. It uses caching appliances (physical or virtual) to provide secure file-based access to an enterprise’s global file repository at local network speeds.
CTERA is another cloud-native solution that provides a global file system and caching at the edge. CTERA’s focus is on transitioning file workloads to cloud without sacrificing security, performance, and control.
NetApp’s leaders recognized that enterprise requirements were changing. To meet these emerging requirements, NetApp introduced the data fabric. Ironically, NetApp is pivoting to become a hybrid and multi-cloud data management services company rather than a seller of networked storage appliances.
NetApp recently acquired Talon FAST and now markets the solution as the NetApp Global File Cache. This product layers the Talon FAST environment on top of an existing NetApp cloud service (Azure NetApp Files, Cloud Volumes ONTAP, Cloud Volumes Service). NetApp deploys its solution as caching software installed on physical Windows servers or Windows Server virtual machines at edge locations. Real-time central file locking is provided through a single ONTAP instance in the cloud.
Dell EMC Isilon has a huge installed base in the enterprise. Dell EMC’s approach to Isilon hybrid cloud is to place Isilon hardware in colocation facilities near the major public cloud facilities and manage the Isilon clusters for its customers. Isilon can also tier cold data to the cloud via CloudPools, an extension of its SmartPools data tiering framework.
Existing HPC solution providers are entering into the enterprise marketplace now. For example, DDN acquired Tintri and IntelliFlash. In doing so, DDN gained storage analytics and proactive support capabilities that enterprises expect from their infrastructure providers. DDN will likely integrate these capabilities with their HPC products, much as HPE did with the InfoSight technology it acquired with Nimble Storage.
Understanding your requirements is foundational
Enterprises are embracing digital transformation. Many digital transformation initiatives are pushing enterprises to consider new file service infrastructures to create new value and carry them into the future. If the COVID-19 pandemic has taught us anything, it is that we do not know what the future will require of us.
Nevertheless, we must act in the present in the face of uncertainty. Speed matters. So does flexibility and optionality. If the legacy filer solution meets your needs, then the fastest path may be a hybrid-cloud implementation with the current provider.
If individual file sizes exceed 16TB, then you will certainly need to look at new solutions architected for the cloud such as Nasuni and Qumulo. If your challenge is providing fast access and file-based collaboration to distributed locations, then a global file system with caching at edge locations may be the best fit. Depending on overall capacity requirements and the number of locations you must support, your short list may include CTERA, Nasuni, Panzura, Qumulo, or NetApp ONTAP and its Global File Cache.
Many solution providers have stepped up to help enterprises create new value from file-based workflows. If you approach the refresh of your file infrastructure with a good understanding of your requirements, you are sure to find a solution that will meet your needs and enable your enterprise cloud transformation.
DCIG will continue to cover developments in enterprise storage. If you haven’t already done so, please sign up for the weekly DCIG Newsletter so that we can keep you informed of these ongoing developments.