Distributed hybrid clouds require modern data management
A gray telephone with a dial stands on a desk. This image suddenly formed in my mind while I was sitting in a discussion group a few weeks ago. The topic under discussion was data management.
It was about the developments that are currently taking place in the use of data. What surprised me was that the focus of most of the participants in the discussion was clearly on the internal use of data. Although the idea of using data for business with external stakeholders is obvious.
The right approach to this also came up in the round: data must be made easy to consume. For example, the logistics department should gain insights from movement data of packages or trucks and react quickly.
One participant in the group suggested that an employee could inform the customer about delays. At that moment, in my mind's eye, this employee reached for the gray telephone. I could literally hear the dial turning.
But it's not just about making data available to a hard-to-scale group of employees - and trusting them to do the right thing with it. Consuming data as easily as possible also means informing customers, dealers and partners directly via web services and mobile services.
A cloud infrastructure is ideal for such applications, but it poses a challenge when it comes to data management: Both internal systems in the local data center and web applications in the cloud access the same data, which must therefore be kept synchronized.
Anyone who wants to use a classic relational database for this cannot meet this challenge. This is because the company would have to add another database in the cloud for the web or mobile app. However, this creates a data silo that has to be permanently compared with the on-premises data in almost real time. Otherwise, incorrect information is the result.
The open source database Apache Cassandra is better suited for this. Cassandra itself uses a distributed architecture, which is why the Java-based, non-relational database optimally supports distributed applications and hybrid systems.
A user sets up a database cluster that connects various clouds and internal data centers. This makes the same data available to both internal systems and any apps in various clouds.
In addition, Cassandra allows companies to add or remove logical data centers in another cloud or in their internal data centers to the cluster at any time - during operation. Cassandra automatically synchronizes data between internal and cloud platforms - virtually in real time.
Therefore, many associate Cassandra with enormous performance and scalability. However, these criteria are not relevant for every application. High availability, on the other hand, is required by every application. Because even when switching to a new app version, a column often has to be inserted into a table.
With a relational database, this table cannot be accessed to add the required column. With Cassandra, on the other hand, this is possible without downtime.
In practice, distributed hybrid clouds are developing. These require easily consumable data management. The large digital corporations faced this challenge first.
Meanwhile, Cassandra and its commercial version, DataStax Enterprise, have been implemented by many top tech players - including Apple, Netflix, Twitter or FedEx. They optimize their business processes, stakeholder engagement, and customer experience using Right Now applications.
The use of the Enterprise variant also ensures them support, training and services for easy administration, operating or monitoring of the databases.
