Forget about Overloading, Here is the Power of Hub

How Hub Integrates Servers to a Centralized System for Data Publishing

Sulung Suganda
Tokopedia Data
3 min readJan 16, 2019

--

Talking about hub, of course it can’t be separated from one of network technologies called hub network. Hub network is one of technologies for connecting multiple computers. The function of hub is also as a center line for distributing a package that sends items into computer network. Basically, this paradigm can be found in our daily life. For example, when we go shopping in e-marketplace or e-commerce, of course, we are required to determine the logistic that we will use for shipping the purchased items, right?

It is clear that the items are not sent directly to buyers. Instead, the items will initially be collected in central office to be checked and recorded. The central office then decides when each item is scheduled to ship, who the courier is, and where the destinations are.

Why we need data-hub ?

We have many partners for marketing purposes. They usually have their own services and we have to feed data to them. At first, every event-data is sent directly by each platform to one or more partners, such as tracking partners. You can imagine if event-data in each platform needs to be published to some partners, we have to update the data in every platform. And if we add a new partner, it will be more complicated, because we should update the platform by using SDK or library that are required by partners. As a result, it makes the application size bigger. That is why we need Data-hub to distribute data using one centralized system.

How data-hub works ?

Data-hub is a centralized system, in which the data from platforms will only be published to one destination. Those data will be managed completely by data-hub. With data-hub, a platform can publish and stop the data from its partners easily, even if we add new partners. And the most important thing is that we can copy and publish the data to many partners directly through data-hub.

Data-hub is not a portable machine, but some servers that work together as a centralized system for user to publish data. Data-hub is designed to handle variety of data. So we build this platform as flexible as possible by using generic method for parsing and formatting the original data. So, what is the main of data-hub?

Data-hub server works based on metadata, in which it always checks metadata out for each data.

Meta-hub (metadata hub)

Meta-hub is an information from data that is managed by data-hub. At first, we have to register the data in meta-hub in order to be completely managed by data-hub. The functions of meta-hub are :

  1. Whitelisted event-data
  2. Data formation
  3. Data standardization
  4. Data transformation
  5. Determine destination partners

Conclusion

Data is important asset, we have to make sure that data is managed well. Data-hub is prepared for handling Volume, Velocity and Variety of data. With data-hub, we can reduce publishing from platform side, so platform will not be burdened to publish many data into partners.

Source: Pixabay

--

--