For companies considering a cloud first strategy, migrating database applications is high on the priority list. Organizations are growing tired of provisioning new servers and buying new storage systems to keep up with database application rollout. They are also tired of the never ending upgrade process the applications insist on as they scale. Applications also go through a lifecycle and it is not uncommon for a popular application to become less so to the point of only being maintained for reference. The problem is the data center architecture can’t adapt to these changing application roles.
The software defined data center is supposed to provide IT with new flexibilities in addressing the needs of applications like databases. The problem though is the storage component of those architectures. Our on demand webinar, “Overcoming the Storage Roadblock toward Data Center Modernization”, discusses how the software data center needs a better software defined storage solution. In this column we discuss how IT can leverage the concepts learned in that webinar to create on-premises database as a service, architectures.
The Problem with Legacy Database Storage Architectures
Typical storage infrastructure limitations force IT to buy a storage system per database application. They are forced to tune that storage system for the specific database use case and they have to buy capabilities assuming the worse case, peak load performance expectations. The result is the organization almost always over buys at the beginning of the project. When the application reaches its pinnacle level of use, the architecture is still sometimes overpowered or other times IT under-projects utilization and they have to do in-production upgrades. Finally, as the use of the application subsides there is no way to “give-back” some of those storage resources.
Is the Cloud an Escape?
Organizations are looking to the cloud as an escape from all this database drudgery. Public and regional providers offer Database as a Service (DBaaS) capabilities that enable the organization to tap into an infrastructure designed specifically for databases. One of the primary aspects of a DBaaS offering is its storage foundation. These providers create a storage infrastructure that is flexible, multi-tenant and can adapt to changing service level needs.
The problem is of course that the organization has to migrate its applications and data to the cloud provider and then “rent” those services and storage for the life of the database, which depending on the application could be lengthy. Renting services for a long time adds up and eventually supersedes the original cost of the on-premises system.
On-premises Database as a Service
An alternative is to bring the database as a service concept, on-premises. The software-defined data center lays the groundwork for this move but the storage layer needs to change in order to truly provide data base as a service to IT’s customers. The storage infrastructure has to be flexible enough to leverage commodity off the shelf server and storage hardware. It also has to perform well so that database applications can deliver the performance users expect. At the same time, the solution needs to provide a quality of service (QoS) capability that enables it to NOT provide full performance to applications that don’t need it, such as when an application is in initial deployment or is sun setting, for example.
The storage software needs to be programmable so that software defined compute and software defined networking can automatically program storage, creating a seamless database workflow as the application goes through its lifecycle.
You can learn all about how storage needs to adapt to become part of the software defined data center by watching our recent on demand webinar “Overcoming the Storage Roadblock toward Data Center Modernization.” With the storage roadblock removed, projects like database as a service become cost effective alternatives to cloud based use cases.