Data Model

What is is a hosted Elasticsearch service with built-in publish/subscribe support for streaming document updates and query results.

It enables you to:

  • Build a blazing fast text search, pub/sub based messaging system, or implement a combination of esoteric filters (fuzzy, geo, terms, range, multiple items),
  • Stream JSON results directly as new data is added or when the original data is updated,
  • Launch in days with our hosted APIs and scale without vendor-lockin by deploying on your choice of cloud provider.

Appbase Architecture
Image 1: architecture overview

We have production users running e-commerce stores, analytics dashboards, feeds, and realtime backends using

There are some catches if you intend to:

  • Model financial or sensitive data - is not ACIDic and doesn’t support multi-document transactions. A good design choice in such a situation would be to use something that supports ACID transactions for storing sensitive data, and use for the data that needs to be searchable in realtime.
  • Perform analytical processing - Being based on Elasticsearch, is designed as an OLTP system although it supports aggregations and queries on data sets of the size of hundreds of gigabytes and even a few terabytes. There are plenty of ideal tools for OLAP use-cases - Amazon Redshift, Google Big Query, Apache Hadoop. can be used in a complementary fashion with any of these for handling online transactions.

Appbase Data Schema

Appbase uses the same data schema as ElasticSearch v5.x and v2.x.

App aka Index

An app in is stored as an index in ElasticSearch. App is the highest level of construct supported by appbase and is equivalent to a database in SQL.


JSON data (aka documents) in an app is logicaly partitioned using types, think of them as namespaces for arranging and querying similar kinds of data.


A document is the actual data stored in an app. Documents are JSON objects, equivalent to records or rows in SQL.

   "user": "data",
   "remember": "something",
   "molala": "yaya",
   "lifeis": 42,
   "recursive": {
   		"how's this": "one nested value"

Visualizing the Data Model

An app can have one or more types and each type is a container for JSON documents. Visually, the arrangement looks like below:

Data Model View

Click on the image above to see the actual data, visualized with data browser.