At Droidcon Berlin, we noticed a lot of questions around databases and that many people weren’t aware of SQLite alternatives and Object-Relational Mappers (ORMs). Therefore, we follow up with an overview of the mobile database landscape.
What is a mobile database?
While Wikipedia defines a mobile database as “either a stationary database that can be connected to by a mobile computing device […] over a mobile network, or a database which is actually stored by the mobile device,” we solely refer to databases that run on the mobile device itself.
Why use a mobile database?
There are some advantages associated with using a mobile database:
- full offline modus for apps that depend on stored data
- frugal on bandwidth for apps that depend on stored data
- stable and predictable performance independent from network availability
- (personal data can be stored with the user, where some say they belong)
What are the advantages and disadvantages of working with SQLite?
The most established mobile database – one could even say the only “established” – mobile database is SQLite. This may be due to SQLite being around since the year 2000 and being embedded with iOS and Android since the beginning. SQLite is a relational database.
In our previous post, we presented advantages and disadvantages of using raw SQLite without any tools as we had gathered them with a group of developers. Based on this, we generated a more generic overview of advantages and disadvantages of using raw SQLite:
What are SQLite alternatives?
There are plenty of SQLite alternatives. If you simply find it unpleasant to write a lot of SQL and boilerplate code, you can use a object abstraction on top of SQLite. This abstraction is usually an ORM (object/relational mapping). But if you want to replace SQLite completely, there are also quite a few alternative databases: Couchbase Lite, Interbase, LevelDB, Oracle Berkeley DB (formerly Oracle's mobile database was "Oracle Database Lite"), Realm, SnappyDB, Sparksee Mobile (graph database, brand-new at the time of this article), SQL Anywhere, SQL Server Compact (discontinued), and UnQLite.
To give you an overview, we have compiled a small comparison table:
|Name||Android / iOS||Type of data stored||Sync Central||Sync P2P||Data level|
|Android / iOS||JSON Documents / NoSQL db||Yes||Yes||Database encryption with SQLCipher|
|Apache 2.0||Embedded / portable db with P2P and central synchronization (sync) support. Secure SSL.||Couchbase|
|Android / iOS||Key-value pairs / NoSQL db||No||No||No||Apache 2.0||Portable lightweight key-value store, NoSQL database.|
|Android / iOS||Relational||Depends on version.||Depends on version.||Depends on version (Lite versus ToGo).||Proprietary||Embeddable SQL database.||Embarcadero|
|Android / iOS||Key-value pairs / NoSQL db||No||No||No||New BSD||Portable lightweight key-value store, NoSQL db, doesn't support indexes, very fast for some use cases; earlier available enchmarks|
from 2011 have been removed unfortunately.
|Android / iOS||Relational and Key-Value-Store||With Oracle Mobile Server||With Oracle Mobile Server||128-bit AES Standard encrytion||Proprietary||Embedded / portable db with P2P and central sync support as well as support for sync with SQLite.||< 1MB||Oracle Corporation|
|Android||Key-value pairs / NoSQL db||No||No||No||Apache 2.0||Portable lightweight key-value store, NoSQL db based on LevelDB.||Nabil|
|Android / iOS||Object Database||No||No||Yes||Proprietary with Apache 2.0 License APIs.||Embedded object db.||Realm Inc.|
|Android / iOS||Relational||Dependant||No||AES-FIPS cipher encryption for full database or selected tables||Proprietary||Embedded / portable db with central snyc support with a|
iOS and Android
|Relational||No||No||No, Use SQLCipher to encrypt SQLite||Public domain||C programming library; probably 90% market share (very personal assumption, 2016).||500KiB||Hwaci|
|Android / iOS||Key-value pairs / document store / NoSQL db||No||No||2-Clause BSD||Portable lightweight embedded db; self-contained C library|
What about NoSQL on Mobile?
NoSQL is a rather large bracket for database approaches that use a data structure that is non-relational. NoSQL databases include key-value stores, document databases, wide-column stores, object databases, and graph databases. NoSQL is generally associated with scalability and performance. However, the scalability of NoSQL approaches on the serverside largely stems from scaling horizontally (adding more servers), which usually is no advantage for a mobile database running on mobile devices.
While NoSQL approaches usually process some operations faster, the particular fit of a particular NoSQL database depends on the use case. Also, the speed of a NoSQL approach may come with less emphasis on reliable data storage (see ACID). So, if your app requires strong data consistency or you handle highly structured data, like e.g. an address book, a relational database may still be a good choice. If, however, you are looking to handle unstructured data or need a lot of flexibility, a NoSQL approach may be more interesting for you.
The general traits of NoSQL approaches that make it worthy for evaluation for mobile:
Object-oriented code in iOS and Android Apps
Developers implement apps in object-oriented languages, meaning an app works with objects. But a relational database works with columns and rows and does not allow storing objects directly. Consequently, objects need to be serialized or reassembled in possibly inefficient way when stored or retrieved. For any database operation this takes time and imposes a natural speed limitation. The right NoSQL approach, e.g. an object database/store, may solve it without complex mappings.
Dynamic requirements of Mobile Apps
After the release of a mobile app, there is often an ongoing live operation process to continually adapt and enhance the app (bug fixes, new features, changes due to changes in the software or legal environment). Implementing changes in a mobile app with a relational client database requires changes to the database schema, which results in additional work. NoSQL databases usually operate without a schema (usually with some enforcing data validation rules). Adding new fields as needed and storing dissimilar data together as necessary means minimal implementing effort for the database. However, again it all depends on the specific database implementation and use case.
Growing need for handling big and / or unstructured data on Mobile
There seems to be the need for handling more and more data (roughly speaking, the amount of data is doubling every 24 months) generally and on mobile. Obviously, on mobile the storage space is still a major concern, which usually omits simply storing every possible data point for potential use. Nevertheless, most NoSQL databases favor speed and store unstructured data. So, for mobile app handling large amounts of (unstructured) data, a NoSQL approach may be the solution.
However, other criteria like size, battery-friendliness or security may also be important factors for choosing a mobile database.
All in all, we hope, we can provide a good general overview of the general status of mobile databases for the summer 2016 with this article.