Chances are you have challenges delivering storage-as-service regardless of the make-up of your data center. Most likely these challenges are a result of exploding data growth, different user data demands driven by a variety of workloads spanning traditional enterprise and scale-out cloud applications, a heterogeneous storage environment (including multiple software management tools), and multiple locations spreading your data center infrastructure over great distances. Add to this scenario a hybrid cloud computing model and the challenges get compounded by the multiple arrays and multiple locations of your public cloud service providers.
The storage industry has been moving in the direction of addressing some if not all of these challenges for some time. NetApp ONTAP, IBM SVC, and others provide heterogeneous storage support but no offering preserves the value propositions of third-party storage. Additionally, no vendor provides the integrated management platform to coordinate the delivery of storage-as-a-service regardless of cloud model—though some startups do offer software and services to construct solutions. A closer look at the storage challenges gives a good indication of what is needed for effective solutions.
While storage virtualization solutions in the recent past offered some promise in minimizing the number of vendor specific management functions, they often became bottlenecks due to their interposing in the data path and duplicating functions already in the arrays like snapshots and replication; these solutions did not leverage the full capabilities of the underlying storage technologies, scale, or offer the means to manage the combined storage as a single entity. More recently, some storage virtualization has moved into hypervisors from VMware and Microsoft, but again the ability to manage and scale on an enterprise-level is missing.
The management challenges resulting from a distributed, multi-vendor storage environment are common to cloud environments, as well as physical and virtual environments, and can result in tensions in the data center from missed project deadlines and a chronic inability to meet service-level agreements. These challenges result from multiple control points, management abstractions, steps in IT processes, and views into storage capacity and usage. Let’s look at each of these challenges briefly:
- Multiple control points: There are too many storage control points within a data center, including separate control points for each product line from a single storage vendor. With most organizations running a multi-vendor storage strategy, the number of control points gets compounded by the number of vendors. Multiple control points mean different syntax and commands for the different storage devices.
- Multiple complex management abstractions: There are also disparate and complex storage management abstractions present because of the different management solutions designed for each array type (e.g. block, file); there is no one standard way to manage it all.
- Multiple provisioning steps: Additionally, a multi-step provisioning process with numerous hand-offs among application, virtualization, server, and storage administrators is commonly required to deliver storage resources to developers. Because the required storage depends on the application, storage resources may need to be some combination of array type increasing the number of steps and staff involved.
- Multiple views: With multiple control points and abstractions, visibility into available capacity, usage, and general health of storage resources is not consistent in the information available and how it is presented to developers and administrators. Data center personnel must create spreadsheets to compile a topology of the storage landscape, which is both time consuming and prone to human error.
As a result of this multiplicity, provisioning storage for users is slow and the operations expense for maintaining multiple management tools, training staff on the different tools, and manually compiling and analyzing storage availability and usage is high.
What if you could simplify your data storage management to get the most out of cloud-era computing?
Cloud computing has been embraced as the next step in the evolution of IT because of the ability to standup new applications and access compute resources on-demand to meet the cost expectations and demands of the global economy. To meet the storage requirements of this new IT model, storage operations need to be agile and efficient on a global scale.
This means that storage operations and management need to evolve to provide:
- Single storage control point for a data center: You need to gain control over exploding storage growth, minimize your storage control points, and sum up your storage with the ability to federate your storage resources over any distance.
- Simple, developer-friendly storage management: For your administrators, you need to be able to equip them with the ability to establish and present cloud storage services that leverage their investment and the advanced capabilities of the underlying storage technology, to control who uses these services, and to monitor and meter usage on a per project, group, and user basis.
- Storage self-service for developers: For your developers, you need to be able to establish classes of storage services backed by policy-based automation that they can self-manage as needed. These services need to encompass primary, backup, and archived storage, as well as security and business continuity needs (e.g. replication and disaster recovery).
- Clear visibility into storage resources: You also need to provide clear visibility into storage resources so administrators know who is using what (and for what purpose), and end-users are aware of their consumption.
Unlike storage virtualization of the not-too-distant past, solving the multiple storage challenges in the cloud model requires a new control path focused on standardizing the build-out and delivery of traditional enterprise-class storage services while achieving the operational simplicity and efficiency of the cloud. Cloud computing requires storage management solutions that harness all of the power of the underlying storage technologies and combine and simplify management and presentation of the storage resources in a user-defined and user-driven services-model.
As always, IT technologies are evolutionary not revolutionary. Some storage management products available today provide some of the storage operations and management capabilities described above. For example, EMC ProSphere offers visibility into EMC-specific storage and is extensible via a REST API. EMC DataBridge will use this REST API to enable EMC infrastructure management products like ProSphere and customer data sources to be combined into new, custom visualizations of storage resources. Additionally, a number of vendors have already embraced and are building to OpenStack which is an open-source, massively scalable cloud operating system that provides a distributed, API-accessible storage platform with support for block- and object-based array types. While all well and good, the market still has a ways to go for it all to come together.