Private Cloud is Dependent on Scalable Storage Architecture
Is your business in the class of businesses that believes that public cloud Infrastructure-as-a-Service (IaaS) poses too high of a risk? If so, then there’s a good likelihood that you are exploring developing private cloud(s) and if that is the case, you will want to seriously consider the role of storage in deployment of your private cloud.
Storage is an interesting animal in today’s digital economy. On one hand, it has become extremely inexpensive to purchase capacity as a commodity, yet it is still expensive to integrate that commodity into a scalable, high-performing and resilient cloud architecture. A key reason for this duplicity is that capacity is but a small aspect of meeting the needs of cloud consumers.
First, one specification cannot satisfy all cloud requirements. A single cloud may have to support a high ratio of write to reads for one application and a high ratio of reads to write for another. Cloud must also offer various application use cases, such as file, object and block access. Meeting all these requirements, while also trying to deliver rapid or self-service provisioning of storage is a challenge for the cloud service providers that do this every day let alone an overloaded IT organization that has the added burden of delivering applications and data to the business.
In addition to the challenges posed by applications that will consume cloud services, storage architecture and engineering skills for building out storage services still lag the available capabilities. It’s no longer enough to be an expert on a particular storage platform. Engineers and architects must understand how to design and manage hierarchical storage across multiple storage vendors’ platforms. Additionally, there is a requirement to understand the networking implications for moving data to and from the multiple layers of storage.
Finally, the storage team needs to deal with an emerging need for more comprehensive governance and security around the storage environment. As design migrates toward shared multi-tenant storage requirements, drives are no longer dedicated to single applications, but house bytes from multiple applications. This is done to deliver optimal retrieval times and increase security of the data, but it dramatically changes how storage participates in a cloud environment. Where data may have used one or two paths previously, it may now be flowing across every available pathway.
So far, we’ve outlined many of the challenges facing businesses deploying a private cloud solution, but the good news is, there’s also some great solutions to help businesses succeed. Converged infrastructures are a great solution for developing scalable cloud solutions. These infrastructures tie together compute, network and storage into an integrated and scalable architecture, taking a lot of the guesswork out of engineering in a piecemeal fashion. Additionally, there are some emerging solutions that offer node-based scalable cloud storage that scale out with the introduction of additional nodes with minimal configuration overhead. These solutions also integrate with existing cloud management systems and simplify administration and monitoring, while providing automation interfaces to foster self-service. Some of these solutions also provide metering, allowing IT to move toward showback/chargeback.
Of note, public cloud is not without its storage challenges either. Sometimes it can be more difficult to troubleshoot performance issues due to storage when using IaaS than when using an internally developed private cloud. However, if you are going to proceed down the route of deploying private cloud(s), then it’s imperative to have appropriately skilled individuals on the team that understand the aforementioned concepts and the needs of the consumers for your cloud.