Graph technology on the rise
Estimated reading time: 3 minutes
We probably use graph technology everyday, when we shop online, spend time on social media or search for something.
You would have participated in 3 use cases for graph analytics, via the simple act of searching on Google and clicking on a result. Graph technology is used by Amazon, Facebook, Google, Alibaba, who all have a combined market capitalisation of over USD2.5 trillion.
Table of contents
So, it goes without saying, graph is huge business. Gartner predicts graph technologies will be used in 80-percent of data and analytics innovations by 2025.
Today, that figure is only 10-percent.
All these simply beg the question: What exactly is graph?
Database technology with rigid structures
According to TigerGraph’s head of solutions engineer in APJ, Ho Chung, database is the ability to to store data in some form or fashion. One form of this, relational databases, have been around for over 50 years.
A simple explanation of this is a platform to store data and organise it so it makes sense, by categorising data across tables and columns, in an Excel-like capability.
If you use spreadsheets frequently and extensively you may be able to appreciate a relational database’s capability to organise data and see relationships between different datasets.
If you have used spreadsheets before, you will know how useful it is to help organise data. You may be able to appreciate a relational database’s capability to organise data and see relationships between different datasets.
Chung said, “But the limitation of relational database is that once you have developed this structure called the schema (simply described as a customer table with all the associated columns), it is difficult to make changes to it.”
Letting data flow like a graph
It is difficult to add, remove, or change due to changes in the business, when a schema is predefined. For example you can’t add new columns or new data (to the schema). This way of modelling the data is rigid and hence, it does not allow for efficient data queries.
Chung explained that about over 15 years ago, NoSQL database technology came around and it allowed us to get away from the strict rigid structures of a schema.
“Instead we just store all the data, regardless if its structured data, unstructured data and throw this all on one platform called Hadoop or NoSQL.”
This solved the problem of perpetually growing amounts of data and the ever growing number of data types, but only for a while.
Graph is the way we naturally think
This database, or way to store and organise information, had to evolve once again.
This was necessary to enable something called deep link analytics, the ability to look at identifying and relating information within a dataset.
This is where TigerGraph comes in with graph database technology. It is able to provide a flexible schema, meaning data is stored any way you want to be able to make changes on the fly.
Another way to look at it is, it provided the ability to model data more intuitively and query it more efficiently.
Native parallel graphs are able to compute deep link features in real-time and at scale. Something like 2 billion calls a day would be a piece of cake, while previous database architectures struggle.
“You will also have high performance on querying the information, just like with previous databases, but with the added notable capability of deep link analytics,” Chung described.
Because of this ability and TigerGraph’s database architecture, it is able to uncover deeper insights and hidden relationships/connections in real-time.
Native parallel graphs are able to compute deep link features in real-time and at scale. Something like 2 billion calls a day would be a piece of cake, while previous database architectures struggle.