In our recently published cloud tiering paper, we reviewed the options for storage-based tiering, gateways and file-based tiering. In this post, I’ll review the challenges with cloud storage gateways to ensure you choose the best, no lock-in path to the cloud for your file and object data. By doing so, you can avoid paying 300% higher ongoing costs in addition to unnecessary initial investments.
Cloud Storage Gateways: A Brief History
In the early days of the cloud, few applications natively spoke cloud storage or object APIs such as AWS S3. Cloud gateways arose to provide a file interface backed by the cloud. The promise was a device with fast disk or SSD to cache hot data while cold data was tiered to the cloud. Cloud gateways promised high performance and the unlimited scale and durability of the cloud with the added benefit of providing access via file interface at multiple locations.
Over the years storage teams found that cloud gateways are useful for collaboration by teams at distinct locations with applications that needed file protocol access to the same data set with modest performance requirements. Cloud gateways could also be used to cache data to small office locations for local access where deploying a full enterprise NAS is an overkill.
An Expensive, Inflexible Path to the Cloud…
Cloud storage gateways commit all data to the cloud for an authoritative copy to be distributed to other locations and keep only “hot data” on premises in cache. This works well for applications that write files and modify infrequently but not for demanding workloads such as database or virtualization. For efficiency, data migrated to the cloud via gateways is stored in a proprietary “block” format. This means the data stored in the cloud is only readable by the cloud gateways, locking in customers and eliminating open access by cloud native tools such as compute, analytics, and AI/ML.
Here’s a brief review of why cloud gateways are not an ideal solution for migrating NAS data to the cloud:
- Additional On-Premises Infrastructure: Cloud storage gateways are typically hardware-based since they serve hot data from the cache. As they act as the primary file server for data that users are actively accessing, this requires performant hardware at each site, increasing costs.
- Duplication of Data in the Cloud: Cloud storage gateways typically put all the data in the cloud and then cache some data locally. So, if you are using a cloud storage gateway for 100TB, then all 100TB of data is in the cloud and a subset of it (maybe 20TB or 30TB) is also cached locally. This means you may need 130TB of infrastructure to house 100TB of data. Depending on the size of the local cache, this duplication of data storage may be larger.
- Licensing Charges to Access Data in the Cloud: You cannot directly access your data in the cloud but must go through the gateway software in the cloud, which means you must pay gateway licensing costs as long as you need your data.
- Lose Benefits of Hybrid Cloud: Organizations want to move to the cloud to innovate and build new workflows using the constantly evolving set of AI and ML tools and services offered by the major cloud providers. But to do that, they need to natively access their data in the cloud. Since cloud gateways simply move legacy storage to the cloud and lock customer data in their proprietary format, they simply do not enable the benefits of hybrid cloud.
So, How Should You Tier to the Cloud?
Where cloud gateways require an entire new storage silo for cloud tiering, Komprise takes a different approach. Working with existing NAS and leveraging analytics to understand data usage and costs, Komprise can help IT teams create intelligent data management policies for cloud tiering, while maintaining open access to data. Data tiered to the cloud is maintained in native format; you don’t need to go through any vendor’s storage system to access data in the cloud and you maintain the flexibility to move to new storage platforms without rehydrating cloud data back to the gateway solution in your data center. Finally, by leveraging the existing NAS, there’s no additional hardware investment required.
Komprise’s commitment to an open-standards format helps to future proof your data, preserving access for your applications today and opening a world of possibilities for future applications and cloud workflows. (See the Gartner 2021 Market Guide for Hybrid Cloud Storage, which includes Komprise.)
Figure 1: Cloud storage gateways vs. Komprise for file-level cloud tiering.
Learn more about the migration path for cloud.
Reviewing the ROI
Komprise is a significantly more cost-effective solution for cloud tiering versus gateways, saving organizations on average 70% annually on storage, backup and replication costs. Here’s the breakdown:
Assuming $700/TB per year of cloud storage gateway licensing costs, cloud storage gateways have 287% higher annual costs than using a file-level data management solution with the cloud.
This is a recurring cost that you pay for over the lifetime of your data.
Figure 2: Cloud storage gateways have ~300% higher annual costs than file-tiering with Komprise.
With cloud gateways, you need to guess which workloads can tolerate reduced performance. Komprise allows you to leverage analytics to determine which data sets need the top performance of their existing NAS infrastructure and then tier the rest to warm and cold storage. You can reduce storage spend even as data volumes continue to grow exponentially, shrink backup needs, and reduce friction for cloud migrations.
To learn more about cloud tiering choices, read our recent white paper.