Contact Us Menu
Filters
Categories: Product Tips

You Think You Know Cloud Storage Gateways? Sync Again

Now that I live back on the east coast, I’m especially thankful for the end of a hot and muggy New York summer. I don’t wish the Time Square Subway Station in August on even my worst enemies. It must be contending to become the 8th circle of hell. It’s just… horrible.

If you spent the summer under a rock, you may have missed that the cloud gateway market was equally hot – with over $100M of investment into storage gateway technology in likely and unlikely places. CTERA secured $25 million in June, our peers Avere, Nasuni, Druva also took investments and even our friends EMC got into the game by acquiring TwinStrata’s block storage gateway for their VMAX division. Facing an increase of customer and investment activity, I’ve learned that it’s very hard for many customers to discern what cloud storage gateway is right for them, so in this blog I’ll try to illustrate the differences.

I’m going to assume that the audience already understands the difference between NAS devices and block storage devices. That said, grab your notebooks and pencils, you’re going to school…

Sync vs. Caching. Herein lies the fundamental difference.

There’s effectively two schools of capacity and namespace management when it comes to the cloud storage gateway market, and the differences have significant impact on customer TCO.

Caching Gateways

Caching gateways are designed to host the authoritative storage volume in a public or private cloud data center while retaining frequently accessed data locally in a storage gateway. Intelligent metadata management enables these systems to present the full dataset (cloud-resident and locally-cached data) all as if it were all local data and buffer caches help to accelerate read and write operations despite using the WAN as a backplane.

Buffer cache

 

Caching gateways have a number of advantages:

  • Archival environments can move data to low-cost cloud IaaS
  • With global locking, many gateways can present a global namespace
  • Buffering enables high performance file storage read-write access

 

To achieve the above, design decisions create some limitations:

  • Buffers are built from expensive SSD technology
  • Internet outages make cloud-resident data unavailable nd through smart snapshot technology, snapshots are synced to the cloud.

 

Sync Gateways

Sync gateways have a number of advantages:Sync Gateway

  • File data is always local, sync gateways withstand internet outages
  • Without SSDs for WAN buffering, sync gateways are very affordable
  • Random read access is often much faster than waiting for WAN I/O

To achieve the above, design decisions create some limitations:

  • Large (100TB+) archives do not benefit from cloud economics
  • Each gateway manages its own volume, locking is not global

 

So… with all of this, how do you know when to use what? Here’s a simple rule of thumb…

  • Sync Gateways are often to replace small-medium branch office storage. Think low-end enterprise NAS systems such as Microsoft Storage Server.
  • Caching Gateways are often used to replace or accelerate enterprise NAS systems such as large scale monolithic or scale-out file systems.

So – as you think of how to evolve your branch and remote office storage architecture, be sure to understand data footprint. The below chart maps the cost of a sync gateway (blue line) approach to branch storage vs. a caching gateway (grey line). As you can see… there’s clear cost advantage when dealing with low data footprint in a branch office.

sync gateway approach to branch storage vs. caching gateway

I’ll expand on this in part II of this blog series, in a post specifically focusing on understanding options for Microsoft server replacement.

Get New CTERA Blog Posts Delivered Directly to Your Inbox
Skip to content