IT managers are aware that not all data requires equal accessibility. In any given system, it's reasonable to assume...
that at least half the data is rarely accessed -- some estimate as much as 80% is needed only occasionally. This scenario creates a nearly ideal environment for a cloud storage gateway appliance: It can help fulfill data access requirements while permitting data to be hosted by a low-cost cloud storage service.
When talking about cloud storage in the context of cloud gateways, it could be in reference to either a public cloud or a private cloud. But for geographically separated tier-three/tier-four cloud services, the discussion almost always focuses on public cloud services. Cloud gateway appliances could still play a role between private cloud data centers, but the use case is different. For the purposes of this buyer's checklist, we'll focus on gateway functionality between corporate data centers and third-party cloud storage providers.
The first factor to consider in examining cloud storage gateway appliances is the application use case: How will it be used? Some questions that might be asked at this stage include the following:
- What type of data will be accessed: large files, small files, streaming video or other?
- Will the data be serving applications with highly random data access requirements or will it likely be repetitive?
- Will the data be in an archive format, such as backup and recovery files?
- Which is more important: cost containment or data access?
- Will the gateway be front-ending tier-four storage (tape)?
Download the "Buyer's Checklist on Cloud Storage Gateways" PDF
These questions will help winnow the field of viable offerings by matching the organization's use case to those in the vendor's "sweet spot." Knowing what type of data will be accessed is central to architecting the offering appropriately. If the data access is highly random, then data-prepositioning won't be nearly as beneficial as it would be for streaming video or large file access, for example.
In the first case, a simple gateway would do while a more full-blown data array would be needed to preposition data. If the data repository will be used as a data protection layer, then gateways with data compression and data deduplication might be beneficial.
Data accessibility will be the next -- pardon the pun -- gating factor for determining which gateway appliances will be best suited to a specific implementation. For example, an implementation that primarily addresses an archive tier will have minimal throughput and capacity requirements. On the other hand, a tier-three repository may regularly access and transfer data from the cloud to the primary data center servers, and would thereby benefit from more bandwidth as well as local caching storage to reduce data transfer.
Questions organizations should address in this regard include the following:
- What are the total throughput requirements?
- Will data access be more or less steady, or will there be significant peaks?
- Will data flow be mostly one-way (i.e., backup volumes) or two-way (i.e., real-time access)?
By knowing the desired access attributes of the offering, evaluators can examine cloud gateway products for those attributes. It will also be helpful in properly sizing and scoping the products before a purchase decision is made. Taking the analysis to an even deeper level, IT managers should consider the nature of the data being accessed, specifically:
- Does data need to be encrypted, either at rest, in flight or both?
- Must the data be retained in an immutable state?
- Are there data access, audit requirements or regulatory compliance issues that must be addressed?
This "Buyer's Checklist on Cloud Storage Gateways" will not only help organizations to better evaluate alternative cloud gateway appliances, but help vendors propose tailored offerings that will maximize users' "bang for their buck."
Table of Contents: Buyer's Checklist on Cloud Storage Gateways
Evaluating cloud storage gateway appliances
The benefits of a cloud gateway setup are compelling. Companies can avoid the logistics of housing massive quantities of data (e.g., power, cooling, floor space) while cloud storage providers offer low-cost, fully managed, high-capacity repositories. Using a cloud gateway appliance, organizations can get the lowest-cost storage profile while meeting data accessibility service-level agreements.
If cloud gateway offerings are all about data access, then data movement capabilities will be at the top of the list regarding product architectural considerations. Some products will use their own data movement software, some will have built-in third-party software and others will rely on (or permit) user implementations.
There are two considerations here. The first is "upstream" to connect and communicate with devices in the central data center, which could be directly to servers, SAN switches, IP networks or other storage controllers. The second is "downstream," or data access from the cloud. Gateway appliances must be able to integrate with the specific infrastructure upstream and retrieve data from the downstream devices, but such considerations may eliminate some offerings that lack specific infrastructure support.
The discussion about fitting in with existing data movement software raises the specter of the most critical dependency, which is the existing infrastructure. In this regard, organizations can either pursue an "open field" offering regardless of vendor or seek a vendor-specific strategy that stays within the parameters of vendors already installed in the data center.
A second major dependency will be file system deployments. Many gateway devices have their own file system included with the product, which must span the cloud storage device. This dictates that the cloud provider support the same gateway devices, which is a major consideration. It may not be necessary to have a unified file system, depending on how the cloud tier is utilized. If it serves separate applications, a separate file system may be the easiest route to implementation. Nevertheless, most gateway appliances allow the deployment of third-party file systems, seamlessly integrating into the overall environment. Either way, a common file system between the gateway and the cloud tier is necessary.
Buyers considering gateway appliances will also find a variety of value-added features that may make a difference. But while the options are there, they may not be necessary. Buyers should verify that they're not paying for something they don't need.
Learn about all of these topics by downloading the “Buyer’s Checklist on Cloud Storage Gateways” PDF found at the top of the page.
BIO: Phil Goodwin is a storage consultant and freelance writer.