Considerations for Hybrid Cloud Storage

You might consider various scenarios for a storage architecture when you deploy a hybrid cloud. Remember that in a hybrid model, some of your resources and assets will be on-premises and some will live in the cloud. Here are some possible scenarios:

Your applications and data are on-premises, and your tier 2 and 3 data is stored in a public cloud.

Some of your applications are in a public cloud, your data is on-premises, and your storage is in a public cloud.

You have a private cloud within your enterprise, and you’re managing a private cloud that’s hosted elsewhere.

Some of your applications are in a public cloud along with your data. Some of your applications and data are on-premises. Your storage is both in the cloud and on-premises.

You get the idea. In a hybrid world, there can be multiple permutations in terms of how you architect your applications, data, and storage. So, here’s what you need to be thinking about in terms of storage as you deploy a hybrid cloud:

Interfaces: To store and retrieve data, your applications need an API that connects your local system to the cloud-based storage system. Users should be able to send data to the cloud storage device and access data from it. You need to ensure that the APIs the cloud provider uses are interoperable with your own, because there are few standards for cloud storage (see Chapter 18 for more on standards). In other words, vendors like to use their own APIs.

According to experts, what users want is a standard like the ubiquitous TCP/IP for the network used across all storage interfaces. However, this may be difficult because each vendor may define its own APIs based on SOAP and REST. So, for the near term, there may be similarities, but vendors won’t be completely interoperable.

Security: Security is always a concern. Make sure security measures are in place when data is transferred between storage and on-premises locations, as well as access-control measures once the data is stored. Files need to be secure while in storage, too.

Reliability: Data integrity is also a piece of the hybrid cloud environment. You need to make sure that your data gets from point A to point B and that it maintains its integrity. Your cloud provider might index your data. Its integrity also needs to remain intact when it’s in storage. For example, if indexes are corrupted, you can lose your data. We talk much more about the why and how of security in Chapter 11.

Business continuity: Planned and even unplanned downtime can cause problems for your business. Your storage provider needs to include snapshots, mirroring, and backups, as well as rapid recovery so that if the provider’s system goes down, you’re covered. You also need to make sure that the right service level agreements (SLAs) are in place (for more on SLAs, see Chapter 17).

Reporting and charge-back: Because cloud storage is a pay-as-you-go model, you need to know what your bill will be at the end of the billing cycle. This will include any transactional charges the provider might charge you as well as storage costs.

Management: In a hybrid cloud environment, if you choose to store some of your data on-premises and some in the cloud, you’ll need to be able to manage the environments together. How will service levels be monitored and managed across these environments? How will you know if there’s a problem with your storage provider? It would be nice to be able to manage all of this together, in one spot, in one single “pane of glass.” However, the industry is not there yet, because it’s continuing to evolve its offerings in this space. See Chapter 4 for more on managing a hybrid cloud environment.

Performance/latency: Once you put your data in the cloud, you are subject to latency (delays that occur when processing data) issues. The questions to ask here are these (which we explore more deeply in the next section):

• How quickly will your applications need data?

• What are the risks if data isn’t available in a reasonable timeframe?

• Will your applications experience time-out and thus problems?

• Does the cloud storage provider match or exceed your network speeds?

• Are there any bottlenecks?

The reality is that combining internal private cloud storage with external cloud storage to look like one storage is difficult. Different vendors have different approaches.

Open source: Rackspace offers what it calls Cloud Files as an open source system for standardizing storage between clouds. It’s doing so under the umbrella of Openstack.org.

Federation: EMC provides its Atmos solution as a software-based solution with services loosely coupled and federated to Atmos servers in the cloud. See Chapter 13 for more information on services architectures and federated approaches.

Gateway approaches: These sit between on-premises storage and cloud storage to translate traditional storage to cloud storage. In many cases, it seems to end users that the storage is a NAS-type storage. We talk more about this in the “Defining the elements of storage” sidebar, earlier in this chapter.