Having enough storage capacity is critical for storage devices. If you run out of capacity, the solution may be as simple as adding a new shelf of drives or as complex as adding a whole new data center. While simpler solutions are preferred, the difference between simple or complex solutions can be whether your system is designed to scale-up or scale-out.
Traditionally, the approach has been to scale-up architecture as data increases. However, this solution has many flaws and limitations. Scale-out storage vendors offer a better way to increase storage quickly without the limitations.
Traditional Scale-Up Storage
Scale-up architecture is the most common block and file storage platform. It contains a pair of controllers and several shelves of drives. If you run out of space, you must add another shelf of drives. However, this method is limited to the scalability limits of the storage controllers.
When you reach the capacity limits of the storage controllers, your only option is to add a new system to sit beside the current one. This becomes more challenging as you must migrate the storage and manage two independent silos. As the company’s data grows, so does the number of new systems to cope with demand. This method becomes difficult to manage and inefficient for resource allocation when deciding where workloads should live.
Innovative Scale-Out Storage Solutions
Scale-out storage offers easier scalability. The architecture can be built on industry standard commodity x86 servers with disk storage tied to each node. Even if the disks are internal or external, only the individual server can access the physical server.
Object storage software is installed on each node which combines all nodes to one cluster in a single storage pool. This means a user or application doesn’t know exactly which node the data resides. Data is presented in a storage container with a Fully Qualified Domain Name (FQDN). The object storage systems manage retrieval, storing, and protection of the data objects and placement across all nodes within a cluster.
Scale-out architecture makes it simple to add more nodes to a cluster and redistribute data across the nodes. Each new node added includes additional disk storage capacity, better performance, as well as more RAM, CPU, and networking resources. This avoids the limitations of traditional scale-up methods.
If you move to a multi-DC architecture, you can add nodes in different geographical areas to extend the cluster. That’s because object storage is geo-aware and policies distribute data to other locations. When a user accesses the storage via the FQDN address, the object storage system will look for the node that will offer the best response time to return data to the user.
For example, a user in L.A. would expect to get faster access from the data center in L.A. than they would from the center in New York. The key is maintaining data consistency between locations to make sure the same data is returned regardless of where the user is located.
As data storage needs increase, consider the move toward scale-out storage options. The scalability helps lower costs, keep it simpler, and offer better resource allocation.
Interested in learning more from professional scale-out storage vendors? Contact Datera for more information.