Resilio helps NetApp customers reliably distribute, ingest, and synchronize files at global scale—from edge to core to cloud—across any location and network.
In the case of NetApp Global File Cache (GFC), Resilio offers a software-only alternative to GFC that fully supports NetApp storage platforms and cloud services, as well as 3rd-party storage. This article compares and contrasts the functionality of NetApp GFC and Resilio Platform for use as a global file cache and synchronization solution.
If you would prefer to see a live demo, please get in touch with us!
NetApp GFC (Edge Cache) Background
In March of 2020, NetApp acquired Talon Storage to expand its arsenal of on-premises file storage and cloud storage solutions. Talon was rebranded to NetApp GFC. As of Nov 1, 2022, NetApp rebranded GFC as BlueXP Edge Cache. We’ll use the NetApp GFC nomenclature in this blog.
NetApp GFC Components
NetApp GFC is a high performance, hybrid cloud edge cache solution based on Windows Server instances hosted on-premises and NetApp cloud storage services running in the public cloud. These include: Cloud Volumes ONTAP, Cloud Volumes Service, Azure NetApp Files, and FSx for NetApp ONTAP on AWS.
There are three components to NetApp GFC:
- Global File Cache Management Server – Hosted in the public cloud.
- Cloud platform support includes Microsoft Azure, Amazon AWS, and Google Cloud (GCP).
- Global File Cache Core – Also hosted in the public cloud.
- Global File Cache Edge – Installed on Windows Server instances on-premises (also called “edge instances”).
All of the NetApp GFC components are centrally managed. The solution helps NetApp customers centralize and consolidate unstructured data in the cloud and provide distributed file access from each remote site. File access is provided via SMB-mounted file shares within a Microsoft DFS namespace.
Resilio Platform Components
Resilio Platform is a software-only, agent-based file delivery and edge cache solution. Resilio agents install directly on endpoints, which can be physical, virtual, or containerized systems. Systems can run a variety of operating systems and file systems (NTFS, ReFS, Ext4, among others). The deployed agents work together in parallel to move and sync files. Agents are centrally managed using the Resilio Platform management console.
The management console provides a single pane of glass to create, monitor, and manage all file delivery operations. Approximately 30K agents can be managed per console instance; each console runs on either Windows or Linux. Console instances can also run on physical, virtual, or containerized servers. The console provides centralized management of all agents and “jobs”. Job types include distribution, consolidation, scripting, and synchronization.
One thing we consistently hear from customers is that they appreciate the visibility and control they gain using Resilio. Jobs are easily created and basically “set it and forget it”. But it’s also easy to set or change bandwidth allocations and track job progress. Scripting and APIs provide automation.
GFC Architectural Comparison
Architecturally speaking, NetApp GFC is a hub-and-spoke solution. The “hub” contains the GFC Management Server and GFC Core. Both of these components are hosted in your choice of the world’s leading public clouds (Azure, AWS, GCP). For cloud storage, GFC requires either NetApp Cloud Volumes ONTAP, Cloud Volumes Service, FSx for ONTAP, or Azure NetApp Files (ANF). Which service you deploy will depend on your cloud platform of choice.
Each branch office (remote site) represents an individual “spoke” where the Global File Cache Edge components (also called GFC edge instances) are installed. These edge instances run Windows Server (NetApp has useful documentation and provisioning scripts for deployment of the edge components). Many customers may already be using Microsoft DFS—and NetApp recommends using the DFS namespace with GFC. (NB: Resilio offers a DFS-friendly solution for DFSR Replacement which is fully compatible with the DFS namespace.)
Unlike a hub-and-spoke architecture, Resilio is peer-to-peer (P2P). This means multiple agents work in parallel to move data. The Resilio solution is adaptive and resilient to the network. One of the lesser known “home run” capabilities of Resilio is bidirectional and N-way sync. Resilio offers fine-grained control over how and where files are replicated and synchronized—in any direction.
Using the Resilio management console, you set this up in a job and away you go. Some customers use a simple one-way push or pull, but two-way bidirectional is popular in site-to-site sync. Remote work and server sync leverage many-to-many. This scenario is ideal when multiple users are updating files; and there was a need to sync the changes from remote locations.
ZGT (Zero Gravity Transport) is a proprietary Resilio capability for UDP-based WAN optimization. Any job can use ZGT to overcome latency and packet loss on WANs. Resilio optimizes how files move in bulk across the network–and is smart about how changes/deltas are sent. Both NetApp GFC and Resilio use deduplication techniques to efficiently transfer file deltas.
Hub-and-Spoke vs Peer-to-Peer
With GFC, one-to-many means that all file changes flow through the hub. When a user is mounted to a file share at a remote site, file changes are efficiently captured and sent to the hub by the GFC edge instance. The GFC Core instance then relays the change out to one or more spokes. GFC offers efficient de-dupe and replication between the spoke and the hub hosted in the cloud.
In the Resilio model, one-to-many or many-to-many is a parallel operation. This means that changes can be sent from one endpoint to multiple other endpoints in parallel—bypassing the hub. In situations where you need to scale distribution or synchronization performance, the Resilio model offers numerous advantages. That said, Resilio can be configured to route all job traffic through a hub, if needed.
The time it takes to replicate a file between 2 endpoints takes about the same time as replicating the same file out to 200 endpoints. This scalability offers predictability: You know precisely how long it will take to sync a given payload over a given connection speed.
But the hub-and-spoke workflow may suffice for use cases that don’t need higher performance, where people work in the office or VPN in. Or where users and apps are physically near the data. Alas, in some implementations, the hub in hub-and-spoke can be a SPOF (single point of failure).
In the GFC model, users should be online with access to their branch office file shares through a local connection or VPN; file shares are mounted via SMB. Once the change is pushed to the hub, the update is pushed out to the other edge instances. Speed and performance will depend on a variety of factors—from file types to file access patterns to bandwidth to hub performance. With GFC, all remote users must be connected to the network and have mounted their file shares. Resilio can be set up to work in a similar workflow.
Data Management
Out of the box, Resilio offers a unified view, where files are managed like they would be through the local file management tools offered on Windows, Mac, and Linux. Windows users manage files in Windows Explorer, Mac users in Mac Finder, and so forth.
NetApp GFC will also give Windows users a familiar view. All users connect to file shares via Microsoft DFS and use their native file management tools as well.
Deployment Considerations
Microsoft DFS
NetApp recommends customers use Microsoft DFS with their GFC deployment. Users must mount file shares and work with files stored on the file shares for GFC to work.
Resilio can be used as a replication solution for global namespace technologies such as Microsoft DFS (Distributed File System) and NFS v4. Resilio replaces DFSR (the DFS Replication service.) In this scenario, Resilio agents are installed on the Windows file servers—and users interact with files and shares as they normally would using DFS. Nothing changes on the client side. Unlike DFSR, all files across all locations can be reliably synchronized — irrespective of location, number of sites, number of servers, file size, or number of files. (Resilio has tested syncing 250 million files per job, although this is not a technical design limit.)
The downside is that, in this scenario, end-user client systems are not running Resilio; so if access over WANs or remote work is required—consider using the Resilio solution for remote work.
Edge Cache using Resilio Selective Sync
Resilio can be used as an edge cache on each remote site to selectively download, store, and sync files. Using transparent selective sync, each Windows Server acts as a storage cache. Users connect to the Windows file server over SMB, as they normally would.
Transparent selective sync gives users the ability to browse files stored in remote locations (e.g., cloud storage) and select which files they want to download. Users have a variety of options to download, partially download, synchronize, and keep files on the device. Files are downloaded, cached locally, synchronized, and purged as needed. Check out more on TSS as global cache here.
Remote Work and Hybrid Work
As stated earlier, Resilio is a good solution for remote work when users need fast and easy access to large files (or many files) from any remote location—including places with spotty internet connectivity, like home networks, and WiFI, cell, or even VSAT. Customers in media and entertainment, sports, engineering, transportation, restaurant services, and other data-intensive industries use Resilio to reliably move their data.
By comparison, if all you need is to collaborate on Office documents—and your users have reliable high-speed access to remote sites—then GFC may be a better fit. GFS will work in situations where users are working in the remote office or have a solid, well-performing VPN. Another good advantage of GFC over Resilio—if your application or use case supports it—is file locking. We’ve heard mixed reports from customers using file locking so this will likely be highly dependent on your use cases, file types, applications, and workflows. The big caveat with GFC is requiring online access to file shares—and having all changes routed through the cloud hub.
Using Resilio Connect, each computer system (Macs, PCs, mobile devices) has the ability to sync and share files of any size in any direction, across any type of network. This is usually one-to-many or many-to-many, depending on your use case. Either way, a big benefit is that remote workers can work offline and sync changes when they’re reconnected.
Online vs Offline Access
As stated earlier, in the NetApp GFC workflow each client mounts the branch office file share hosted on an edge instance (a physical or virtual Windows file server) via SMB. The user opens the file directly off the share; when changes are made, the GFC edge instance communicates with the GFC management server hosted in the cloud and efficiently replicates the delta.
Each edge instance caches “active data sets” in remote offices. When a file changes in one location (on one file share), the change is efficiently replicated “up” to the hub in the cloud. From there, the change is trickled out to each of the spokes—which are all of the other edge instances hosted in various remote locations (usually branch offices). Most Microsoft enterprises will likely have all of their file shares linked up in a DFS namespace, and managed through Active Directory. For GFC to work, users need to be online with cloud access.
Using Resilio on each endpoint, the network connection can fail or be intermittent, and the user can work offline. In some cases, multiple users running a Resilio agent may be connected to a remote network and need to share data with other users on the network. Resilio is really really good at finding a way to get your files where they need to be—within a given timeframe.
Agents (which can be read-write or read-only) reliably transfer and sync files without having to access a centralized server, centralized storage infrastructure, or without having to go through a central hub. In the Resilio model, there is no single point of failure (SPOF). If an agent goes down or offline, other agents dynamically route around the failure to find other peers.
On-Premises, Hybrid Cloud, or Public Cloud Native
Both solutions support hybrid multicloud in public clouds (AWS, GCP, Azure). Resilio also supports other cloud platforms, as well as any S3-compatible cloud object storage. Both solutions support NetApp cloud volumes ONTAP and ANF — but NetApp requires this in the cloud where Resilio does not. Resilio runs on virtual machines running your OS of choice—hosted in your MSP of choice.
Resilio operating system support includes Linux, Windows, macOS Server, BSD, and a variety of 3rd-party NAS solutions where Resilio runs natively on the NAS. Besides being cross-platform, another advantage of Resilio Platform is that you can use any type of storage out at the edge, at the core, or in the cloud—including NetApp. This could be direct-attached, SSDs, JBODs, NAS, or SAN. Anything.
Active-Active vs Active-Passive High Availability
For disaster recovery and other cases requiring high availability, Resilio makes it extremely easy to deploy active-active HA. Because Resilio offers low-latency, real-time, omni-directional replication and synchronization—all files can be updated within seconds. Customers report recovery times (RTOs) in minutes and recovery points (RPOs) in seconds.
With DFS in the picture, site to site failover and redirecting users to the active site is greatly simplified. For VDI, some customers use front-end load-balancing like Citrix load balancing services to direct users to the active site based on policies. The key enabler for active-active HA is syncing files fast and in real-time.
For NetApp GFC, you’ll have to follow the best practices for HA. Some NetApp customers may use SnapMirror to set up a one-way replication schedule to run every 15 minutes or so. That equates to a 15 minute RPO. Another consideration is that this is limited to one-way replication between at most 2 sites (endpoints) concurrently. If there are multiple cloud instances running ONTAP, you’ve got to consider replicating across multiple cloud instances for HA.
According to page 46 of the NetApp GFC user guide, NetApp recommends using the HA capabilities in the hypervisor. “In case a GFC core instance fails, and can’t recovered in time, you can ‘replace’ the failed GFC core instance with a ‘cold’ standby instance by either changing the IP address of the ‘cold’ standby instance or updating the DNS record associated with the edge-to-core association (i.e. IP address, Cloud Fabric ID configured on the edge).”
Summary
Both solutions support leading public cloud platforms but Resilio also supports a broad range of storage solutions and can be deployed as a hybrid cloud, completely on-premises, or natively in the public cloud. As a peer-to-peer solution, Resilio offers a push and pull edge cache solution that is well suited to widely distributed deployments requiring fast access to files, real-time replication and/or advanced distribution or synchronization scenarios. Resilio excels at reliably moving bulk payloads of data across a diversity of networks (from poor quality low bandwidth to high bandwidth).
Resilio gives both NetApp and non-NetApp customers the flexibility and synchronization speed to replicate files of any size and type in real-time, for a variety of workloads and workflows. You can replicate files in real-time in any direction—using any device (physical or virtual), running any operating system, and using any type of storage system (block, file, object) from any vendor.
To learn more about the Resilio solution, or discuss pricing, please schedule a demo to get in touch with us! In the meantime, please stay tuned for Part 2 of Resilio as a NetApp GFC alternative.
