Introduction

 

Distributed teams and collaborating together on shared files hasn’t lost any importance in the past few years. It will become even more important over the next few years as many applications and virtual desktops will be moving to the cloud and will live in a hybrid model. Inevitably, file collaboration will follow that same trajectory: some users will keep working in an on premise environment, but other team members will be using a thin client and work in the cloud. However, the requirement that teams collaborate on the same shared files without stepping on each others’ feet will still be a hard requirement. In this post, we’ll look at an option that enables this exact use case.

 

Peer Software’s Approach

 

PeerLink by Peer Software is a file collaboration solution which is designed to integrate with NetApp storage solutions. PeerLink was created to provide distributed teams with a fast and efficient way to collaborate on shared files. PeerLink combines an enterprise-class, real-time synchronization engine with distributed file locking. Its real-time synchronization engine ensures that the same data exists on all participating NetApp systems, regardless of where changes may occur. The file locking component prevents users from accessing files that are currently being worked on by another user at any location.

 

Solution Architecture & Benefits

 

The PeerLink solution integrates with both NetApp Clustered Data ONTAP and 7-Mode systems. This also includes NetApp Private Storage, a near-cloud solution, as well as Cloud ONTAP, an in-cloud virtual NetApp storage system for the Amazon Web Services. Let’s have a quick look at the architecture:

 

peerlink_arch.png

 

In each site (data centers and clouds), the PeerLink Agent is deployed. In the main data center, in this case Germany, the PeerLink Hub is deployed. These two components keep track of global locking and replication. They do not sit in the data path, but rather integrate with NetApp’s FPolicy API.

 

Distributed shared file system & global locking

All users in this diagram can view and edit the same data, regardless of their location. If a user creates a new file, e.g., in Germany, this file will be made available in all other locations. If another user updates the file, e.g., inside the AWS Cloud in Singapore, this change is propagated in a very efficient, file block based way to all locations. However, while a user edits a file, e.g., in the UK, this file gets locked at all other locations, so that we can make sure that users aren’t interfering with each other. This prevents conflicting versions of files across locations.

 

WAN-optimized with a file block-based replication engine between sites

When a user in a location changes a 2GB file, it is not feasible to transfer the entire file to other locations each time the user hits the save button. Therefore, PeerLink listens to the NetApp controller and implements a file block-aware replication engine that only replicates changes on very small block increments. This allows synchronization of updates to large files across multiple locations to be extremely fast and efficient over a WAN.

 

Performance through local copies

Holding files only in a central location is not feasible, as opening would require a complete transfer of the file from the central location to the remote location. Imagine opening a 1GB file that sits in Germany from Singapore. Instead, PeerLink puts copies of required files at each location to allow all users to work with files at LAN speeds, regardless of file size.

 

Summary

 

Peer Software’s PeerLink and NetApp’s Data ONTAP enable distributed teams to work in a shared file system across multiple locations with global locking. Peer Software’s efficient, WAN-optimized replication engine, paired with NetApp’s world-class storage efficiencies provide a scalable solution which can span multiple data centers, as well as into the public cloud.

 

For more information about Peer Software’s PeerLink, please visit their website.

 

Follow me on Twitter: @clemenssiebler

Clemens_Siebler