BangDB 2.0 Architecture

High level

BangDB comes with lots of elements converged into it. We have to look at it from convergence point of view before we go into details. Here is the high level view of BangDB. It indicates the community and enterprise angles clearly and at the same time explains the different layers as well


Lowest layer is about data ingestion. BangDB implements set of data ingestion agents. Apart from that it can also use the protocols like SMTP, Bacnet protocol and directly ingest data if access is provided. The log file ingestion mechanism may use IE to do the transformation in automated manner if enabled. Most of these elements come with Enterprise license

The processing layer is where data is processed. Please note that the data is not yet on the file system and it's analysed much before that. Convergence allows BangDB to avoid network hops and also hitting the file system before analysis, this reduces latency to a level which can't be matched with siloed system

Further, Access and integration layer allow BangDB to connect with ecosystem, such as hadoop, RDB etc. This is useful when we process large amount of data mostly for real time and then archive to batch system like hadoop.

Finally, The dashboard layer, which allows users to build applications in simple manner and publish them for deployment or usage. This can be provisioned on any cloud and deployed within few minutes

Under the hood

Now let's go bit under the hood and see the components. Here is the high level view of the stack or component layer as they stack up


The stacks are self explanatory, however, we must appreciate that in the converged model, we have most of the necessary elements/components or pieces packed in a single box or in a single process, and then that becomes the unit of compute. This unit of compute is what scalable in horizontal manner in order to process huge amount of data.

Therefore, while there is a fabric from high level, within a node as well there is similar fabric to ensure the processing efficiency. More on these in separate topic of distribution etc.