While writing our last blog on NetApp GFC, NetApp rebranded its global file cache (GFC) software. The new names, NetApp Global File Cache and Cloud Volumes Edge Cache, are packaged and placed under a data Mobility theme in the new NetApp BlueXP Edge Cache product suite.
Since the updates seem to be a marketing-only re-brand, we’ll stick with the old product naming scheme: NetApp GFC.
To recap, Resilio is excited to be working closely with NetApp as a strategic solution partner. Resilio helps NetApp customers move their data faster globally to support the most demanding high performance data movement scenarios for file ingest, synchronization, and distribution, among other solutions.
But we also have a flexible alternative to NetApp GFC, as covered in our first blog, NetApp Global File Cache Alternative; in that blog we provided an overview on NetApp GFC and compared the hub-and-spoke architecture to peer-to-peer. We also covered a few deployment considerations.
In this blog, we’ll go deeper on the challenges and advantages of each solution. Finally, we’ll compare NetApp GFC head to head with Resilio Connect.
Traditional Edge Cache Challenges
When unstructured data is widely distributed across multiple locations, there are many ways to cache files—from edge cache network solutions to global file systems to CDNs to storage file system solutions like NetApp GFC.
Through a global file cache edge instance, GFC improves performance to an active dataset over SMB file shares. Thus, end users must have access to their file shares and the “active” file system data must be made available through a DFS namespace; all centralized data will need to be stored in the cloud behind NetApp cloud volumes, ANF, or a similar NetApp cloud storage service.
With GFC, the problem then becomes how to keep all of that widely distributed data rapidly synchronized in real-time across all edge locations: not just remote offices but also home offices and end users’ workstations or mobile devices. This is where Resilio comes in. Resilio gives you a cross-platform, software-only solution to cache, update, and sync files directly in any direction, at global scale. Using NetApp storage; or any storage.
The NetApp GFC model relies on:
- Remote SMB file shares tied to centralized data stored in a hub-and-spoke hybrid cloud: Replication performance will always be limited by the hub; the centralized file services stored in the cloud. In NetApp GFC (and other hub-and-spoke solutions like Nasuni) your centralized data is stored in the cloud. (Using Resilio, you could be all on-prem or in the cloud.). With hub-and-spoke, all file updates need to pass through the hub and trickle out to each edge instance. The more distributed your end users and remote sites, the more challenging it will be to maintain high performance; avoiding SPOFs also becomes more challenging as well. You can make the edge and the hub faster: using SSDs or all-flash arrays may help in terms of IOPS, but no matter how fast and efficiently you handle disk IO, the hub will always be a bottleneck when synchronizing files across more than 2 endpoints.
- Caching at the edge but physical consolidation at the core: Moving data closer to end users and applications is great in practice—if you can manage the data and keep data in sync across all locations. GFC does simplify managing the distributed data, and the DFS namespace makes it easier to access. One consideration is that end users may not always have access to their remote file shares. A second is performance: consider the round-trip time it takes to synchronize many millions of files across multiple remote locations that all must route through the GFC core. It’s like follow the sun with a single stop at the moon. CDNs or smart caching with DNS won’t get around the hub bottleneck either.
- Smart dedupe for large files or diverse payloads. NetApp GFC utilizes a smart dedupe method to keep unwanted traffic off of the network. Yet, synchronization of large files over 1GB may be problematic or cause IO storms. But if your workload and file sizes are smaller (or the projects contain smaller files), then GFC may be a good fit. Please note that Resilio Connect has no limits on file sizes. Or network bandwidth or network capability.
- NetApp only storage options: NetApp has a broad range of cloud storage and on-prem offerings. And some customers may be fine with this. Other customers may need to deploy multi-vendor storage solutions, like an on-prem NetApp shop that happens to using a non-NetApp cloud storage offering; or a team with one project that has different storage. Resilio works with NetApp on-prem and in the cloud; or any other storage type (file, block, object; DAS, NAS, SAN).
- Complex high availability for disaster recovery: The hub is a single point of failure. That means if the hub goes down, all other remote sites go down. Designing ways around this increases complexity and jeopardizes meeting SLAs for availability. Meeting RTOs and RTOs will also depend on how many core and edge sites are deployed. The more sites the higher the complexity to achieve high availability with NetApp GFC. In Resilio’s case, more sites does not add complexity. In fact, the more endpoints working on replication in parallel equates to higher performance. Morever, there is never a single point of failure in the Resilio model.
- Networking and connectivity being always on and users being online 100% of the time. With NetApp GFC, remote users connecting over highly reliable, high-speed broadband may be OK with occasional hiccup. Less reliable networks may cause more serious problems (i.e., WiFI, cable, LTE/5G, etc.). With NetApp GFC end users must not only be connected to their remote office file shares and have access to the cloud–but they have to work across them. Than means opening, editing, and closing files directly from the file share. SMB file shares are made available through a Microsoft Windows Server instance or multiple instances stored at each location; ideally, the shares will be presented through a DFS namespace and have continuous access to NetApp cloud storage (a high performance, pay-as-you-go subscription service offered by NetApp).
Resilio Advantages over NetApp GFC
- Offline work! Resilio handles unreliable networks and connectivity problems without end user intervention. For some use cases, the ability to work offline will be a requirement. Resilio enables a high degree of flexibility for offline access. Users can work offline when their workstation is offline, or when the network at the remote office or main on-prem or cloud facility is unavailable, and still work and share files with nearby peers. When back online, Resilio can re-sync or synchronize anywhere on-demand or on a schedule. This really comes into play when you lose connectivity between the end user and the branch office or between the remote office and the cloud. Once reconnected, Resilio can update all the peers on the local network. In edge computing or other use cases with poor connectivity, it’s nice to have the flexibility to update nearby peers on the local network without having to connect to the hub to see or send changes.
- Anywhere file delivery architecture to replicate and sync files in real-time in any direction: Resilio gives companies a reliable solution for bidirectional sync and N-way sync. Distribute, consolidate, synchronize, and automate file delivery in any direction: one-to-many, many-to-many, or many-to-one. (GFC, by contrast, is 1-way only.) Resilio is able to bypass the hub, when needed; the big enabler to this is peer-to-peer. (Hub-and-spoke is built on a client-server model. Here’s a summary of the difference.)
- No vendor lock-in: Resilio is cross-platform (and runs on Windows, Linux, macOS, Android, FreeBSD and a few NAS systems, too.) You can use any storage, any OS, and any cloud. In terms of storage systems, Resilio supports pretty much any type of storage (file, block, object—from any vendor) both on-premises and/or in the public cloud. There are no dependencies on Windows file servers and NetApp storage—although both are fully supported. Resilio supports a diversity of workflows and operating systems.
- Turnkey data migration: Through a simple check-box, Resilio enables replicating NTFS permissions and other metadata to/from any ONTAP system as well as other platforms. Need to migrate from Isilon to NetApp? Migrate from DDN to NetApp? Or to another cloud storage solution? No problem. Resilio has you covered. Check the box for migrating NTFS permissions and other metadata with files in the Resilio management console.
- Any payload (vs. soft limits on file sizes): Files of varying sizes and types or file systems with 250+ millions of files per directory can be synchronized at global scale. NetApp GFC scales but has soft limits on file size. We’ve heard that GFC may create a “streaming storm” on file sizes over 1GB.
- Simplicity (vs complexity): Resilio is easy to use on Windows, macOS, Linux–and any supported OS. Files are managed just like they are in Mac Finder or Windows Explorer. There’s a low learning curve for end-users. And no dependencies on end-users to find and mount network shares to gain access to files. Users can work online or offline—and be located anywhere–across any network.
- Overcoming data gravity and latency: Resilio reliably moves data across any type of IP network and has the flexibility to distribute and sync files in any direction: from bidirectional to multi-directional (one-to-many, many-to-one, many-to-many) both in parallel and in real-time. A change that happens on server A in site 1, can be directly sent to all other servers within the site—or to any other site—without having to go through a central hub. Combined with Resilio’s Zero Gravity Transport (ZGT) you can move files at full speed across any WAN or high latency network.
- Cost transparency (vs hidden costs): NetApp offers a great GFC product at a low up-front cost. So does Resilio. Both GFC and Resilio are easy to provision and manage. But with GFC, make sure to calculate the ongoing operational costs of running, managing, and replicating the NetApp cloud storage environment. Things like egress add up for both solutions but Resilio does offer ways to minimize the egress traffic. Resilio can pin traffic to a network in the cloud (i.e., keep certain traffic on the cloud provider’s network) as well as use transparent selective sync to efficiently download files to minimize egress traffic. There are lot’s of ways around egress traffic with Resilio.
- Active-Active high availability across all locations: Resilio eliminates single points of failures and protects files across all sites and endpoints. In the GFC model, each edge filer pulls down changes on-demand. This creates a delay in keeping all head-end and edge filers in sync. Unlike GFC, Resilio can reduce RTOs/RPOs, support active-active failover, and keep all endpoints in sync within seconds. GFC pulls down changes to the edge on-demand so when you fail-over you have to start seeding all that data across all of the edge instance VMs.
- High performance scale-out architecture vs client-server: NetApp cloud storage does scale-out. But in the context of global file cache (or edge caching in general) the bottleneck will likely not be storage IO or throughput. The bottleneck is file replication: sending all the new files and file changes back through the hub in the cloud. Resilio enables scale-out file delivery and sync to send updates in real-time to any number of remote target destinations, concurrently across all endpoints.
NetApp Global File Cache – Other Considerations
NetApp GFC is good for NetApp branch office consolidation and intelligent file caching of Microsoft Windows file servers in deployments where connectivity at the edge is always available. End users must remain connected to both SMB file shares, ideally the DFS namespace, and also centralized data hosted in the cloud on NetApp cloud storage such as NetApp cloud volumes ONTAP, the cloud volumes service, NetApp FSx for AWS, or Azure NetApp Files.
Another positive for NetApp GFC is that, if your use case supports it, file locking is also offered for online users connected to Windows file shares. Yet as stated earlier, any change made to a file needs to first be sent to NetApp GFC management instance hosted in the cloud. File locking may also depend on the application workflow and locking mechanisms offered by the ISV (e.g., CAD provider or another app).
Having a way to prevent other users from accessing files is tricky business. File locking should be piloted ahead of time to see if it’s right for your company. And there may also be a performance hit. According to one GFC expert, file locking may introduce some latency—impacting performance.
For data management best practices, NetApp recommends keeping shared projects under 500MBs or copying (staging) projects to edge instances in remote locations ahead of time (i.e., before users will be working on the project). Users should work directly from mounted file shares to access and modify files—and not copy files locally to their desktop machines. You can stay abreast of What’s New in GFC on the NetApp web site.
If data sets are kept relatively small, NetApp GFC can help minimize “chatty” traffic on the network. This becomes a consideration across WANs with latency and precious bandwidth. I’d recommend researching what file types work best for GFC in the context of your workflow. Active file changes will be replicated between each edge instance (spoke) and the cloud store (hub), before being trickled out to other edge instances in the deployment.
In summary, NetApp global file cache addresses use cases such as:
- Efficient file access for users logging on within branch offices to the cloud for CAD and other dedupe-friendly file types that may (or may not) require file locking. NetApp recommends projects be kept under 500MBs for active files. If required, VPN access to each branch office is another consideration, which may impact performance.
- File server or file system consolidation: Companies can logically consolidate file servers through DFS—and physically consolidate their enterprise data in the cloud with NetApp GFC and NetApp cloud storage. Performance will depend on a number of factors, from file access patterns to file size to payload capacity traversing the network. Below we’ll discuss some of the architectural challenges with hub-and-spoke deployments.
- Read-mostly (or read-only) VDI deployments where VDI profiles are small and files cached in remote sites to minimize replication and synchronization traffic. Any change made to the VM has to be pushed up to the GFC hub in the cloud before it can be propagated out to other VMs. If many writes/updates are happening in your VDI environment, that could impact performance and/or availability. Alas, platforms like Microsoft FSLogix, as far as I know, do not support read-only modes.
Where Resilio Connect Wins
Like NetApp GFC, Resilio works with the DFS namespace but in a different replication model (as a DFSR replacement) or you can deploy Resilio directly on all endpoints. In the latter scenario, Resilio is ideal for flexible local caching for Azure (or another cloud provider) for scenarios like remote work.
In either case, the big picture benefit of Resilio is that your files are where you need them—no matter how widely distributed your end users, applications, and sites are—in predictable time frames (i.e., when you need them). Resilio keeps files in sync within sub-5-seconds of a change across up to many thousands of endpoints. Changes can be routed a variety of ways: directly to other peers in the system (or optionally through a hub); across any type of network. Customers have fine grained control of how you move and sync data with Resilio—from one- and two-way to full on N-way.
If you need a cross-platform, storage-independent global file cache solution that is easy to use and supports any IT or cloud infrastructure, Resilio provides tremendous flexibility. You can choose from any cloud platform, or can deploy all on-prem, hybrid multicloud, or all cloud. Resilio is fully compatible with NetApp storage and cloud services and may also be used with non-NetApp cloud or on-prem storage.
There are no limits on file sizes, and Resilio has validated synchronizing up to 250+ million files per job; although this is not a hard design limit.
Resilio is well-suited to push/pull and real-time replication scenarios, such as:
- Server sync, content replication, and DFS replication: Resilio offers flexible real-time replication across sites and cloud regions—and NetApp platforms. One- and two-way bidirectional, one-to-many, many-to-one, or N-way synchronization are enabled.
- VDI profile sync: Resilio speeds up time-to-desktop by replicating changes in any direction across multiple VMs concurrently. An added benefit is active-active high availability and faster failover.
- Multi-cloud or intra-cloud cross-region replication (CRR). Use any on-prem IT infrastructure with any cloud and any type of cloud storage.
- Complement SnapMirror with real-time sync. Resilio one-to-many, many-to-one, and many-to-many sync scenarios across any type of NetApp storage.
- Data migration across storage platforms. Resilio migrates files and permissions to/from any storage platform.
Reasons to Evaluate Resilio as a NetApp GFC Alternative
- No Single Point of Failure. By design, there is no single point of failure in the peer-to-peer Resilio model. Routing is dynamic—to route transfers around failures.
- Files change often at the remote sites. Resilio syncs changes in real-time; if users are offline, changes can be synchronized among available peers and synchronized to other peers as they become available.
- Files may be of any size but also large files over 1GB in size.
- Bidirectional or multi-directional replication and/or synchronization needed at the core or edge — where a hub-and-spoke is the bottleneck.
- Non-NetApp storage any storage solution needs to be supported.
- When operating systems other than Windows need to be supported.
- Multidirectional sync – send changes anywhere to anyone bypassing the cloud.
- Speed matters – Resilio scales performance to over 10 Gbps per system.
- VDI profile sync – with larger profiles and requiring bidirectional, many-to-many, many-to-one.
- Remote work use cases where users need to sync or share data directionally with other users or groups as fast as possible, or needing to collaborate on files over 1GB in size.
|Capability||NetApp GFC||Resilio Connect|
|File size limits||Practical limits of 1GB per file but no hard limit||No limits|
|Multi-directional file replication and sync||One-way, one-to-many via hub and trickle||One-way, two-way, many-to-one, one-to-many, many-to-many (N-way)|
|Selective Sync||Serialized on-demand only; all changes sync through hub (edge GFC change > sync to hub > changes trickle to edge GFC instances)||On-demand push or pull via TSS; or automatic N-way sync|
|Cloud instance replication||One-way scheduled via SnapMirror||Real-time One- , two- , N:1, 1:N, or N-way via Resilio Connect|
|Namespace for Failover||DFS-N||DFS-N|
|Real-time||Yes – But changes flow through a hub and then out to edge sites||Yes – Changes flow in any direction to any endpoint in parallel|
|Workflow requirements||All files must be stored and accessed online via DFS mounted file shares||Files can be stored and accessed online or offline using both DFS file shares and/or local storage or cloud storage|
|Device & OS support||Windows Server only||Windows Server, Windows desktop, macOS Server, Linux, BSD, Android, and a variety of NAS|
|Virtualization support||Physical and virtual machines (.OVA and other formats)||Physical and virtual machines (.OVA and other formats) or containers|
|Storage requirements||Cloud Manager requires Cloud Volumes ONTAP single node or HA pair in Azure or AWSLegacy deployments may use Cloud Volumes ONTAP or Cloud Volumes Service on Google Cloud, Azure NetApp Files, or Amazon FSx for ONTAP||Any storage—any vendor: DAS, NAS, SAN, or Cloud Storage (Object and File Storage)|
|Cloud support||Azure, AWS, Google GCP||Azure, AWS, Google GCP, Oracle, Backblaze, Wasabi, and partner MSPs.|
|WAN Optimization||Dedupe and compression delta difference||Dedupe and compression delta difference. ZGT for bulk data transport optimization over UDP.|
|Automation||Scripting and APIs||Scripting and APIs|
Both solutions work with a variety of cloud environments but Resilio also supports any type of storage system on-premises and in the public cloud. For those customers that need to remain all on-premises, Resilio has you covered. And you can extend to any cloud at any time, if needed.
Resilio gives 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!