Overcoming Bottlenecks with a Faster Database

The cloud database is a relatively new concept of managing databases as-a-service on the cloud. The main idea behind this concept is to free users from all the demanding tasks associated with managing and operating databases. This includes making the database highly available, guaranteeing data durability and solid performance, and managing backups. And, according to Rackspace, cloud-based databases deliver a faster performance with the elimination of performance bottlenecks and built-in data replication.

“More advanced cloud database services completely virtualize the database’s underlying architecture from the users,” explained Yiftach Shoolman, co-founder and CTO of Garantia Data, a provider of advanced in-memory NoSQL cloud services. “For example, the users have no clue how many nodes are processing their requests, where these nodes are located, which storage devices are used to store their dataset. In most cases, advanced cloud databases provide a single DNS address for the users to access their database resources plus some access control mechanisms, leaving the users to deal with database schema design and with managing database requests. These services completely eliminate the need for DBAs, making the app development process faster and cheaper.”

An efficient cloud database service is usually built over a multi-tenant architecture, managing multiple databases of different users over the same cloud instances, usually in a highly secured manner, Shoolman added.

“If a multi-tenant architecture is not built properly, there is a great chance it will suffer from many processing bottlenecks such as CPU, storage I/O, memory and network. These usually lead to higher response time and connectivity issues, leaving the users frustrated,” he said.

However, by turning to cloud-based databases, it is possible to avoid some of the bottlenecks that occur in other, more traditional databases.

“First, you avoid IT bottlenecks by just being able to quickly requisition a DB and get a DB up and running,” said Dave Messinger with Cloudspokes. “Secondly, most database-as-a-service provides all the hardware and virus monitoring. They also handle all the patching of hardware and software. They provide superior monitoring and usually great APIs for you to monitor your database including superior SLAs. I also don’t have to worry about having DBAs and can focus on proper database schema design and code rather than worry about how to set up infrastructure.”

Databases based in the cloud businesses are able to cope with increasing demands regarding the amount of managed data and the number of concurrent users accessing the system, Mike Hummel, CEO with ParStream
pointed out. “The combination of an easily and reliably scaling infrastructure with a cloud database offering that scales in a similar manner does make handling of such scenarios a breeze,” he said.

A scalable fully managed hosting solution helps database performance because it gives the customers the ability to scale on an infrastructure level in very short timeframes and at a predictable cost, Hummel added. “Combining that with an equally scalable and failure resilient database offering that is able to scale with the underlying infrastructure you get a really compelling package.”

Cloud database architects should know exactly which cloud instances and storage devices will avoid bottlenecks, said Shoolman. This is not only a theoretical process, but rather a process that requires days and months of intensive testing in-order to reach the optimal architecture.

The cloud database improves on traditional database offerings and a host provider should be able to easily catch up with the latest cloud offering to guarantee the best service for its users.

Sue Poremba is a freelance writer focusing primarily on security and technology issues and occasionally blogs for cloud service provider Rackspace Hosting.