A graph database is a data management system software. The building blocks are vertices and edges. To put it in a more familiar context, a relational database is also a data management software in which the building blocks are tables. Both require loading data into the software and using a query language or APIs to access the data.
Relational databases boomed in the 1980s. Many commercial companies (i.e. Oracle, Ingres, IBM) backed the relational model (tabular organization) of data management. In that era, the main data management need was to generate reports.
Graph databases didn’t see a greater advantage over relational databases until recent years, when frequent schema changes, managing explosives volume of data, real-time query response time, and more intelligent data activation requirements make people realize the advantages of the graph model.
There are commercial software companies backing this model for many years, including TigerGraph (formerly named GraphSQL), Neo4j, and DataStax. The technology is disrupting many areas, such as supply chain management, e-commerce recommendations, security, fraud detection, utility power grid scheduling, knowledge graph for AI applications, analytical queries on blockchain general ledger data, and many other areas in advanced data analytics. For a comprehensive description, please see this page and this page.
Here, we discuss the major advantages of using graph databases from a data management point of view.
This means very clear, explicit semantics for each query you write. There are no hidden assumptions, such as relational SQL where you have to know how the tables in the FROM clause will implicitly form cartesian products. Besides ease-of-use, such as regular path pattern matching, accumulator concepts allows fine control to keep mid-way query state in-place of the data.
They have superior performance for querying related data, big or small. A native graph has the so-called index-free adjacency property, where each vertex maintains its neighbor vertices information only, no global index about vertex connections exists. This makes native graph exhibit constant performance while data size grows. The performance is constant since that traversing a given vertex’s neighbor nodes via edges are independent of the graph size. It never needs to load or touch unrelated data for a given query. They’re an excellent solution for real-time big data analytical queries where data size grows rapidly.
Graph databases solve problems that are both impractical and practical for relational queries. Examples include iterative algorithms such as PageRank, gradient descent, and other data mining and machine learning algorithms. Research has proved that some graph query languages are Turing complete, meaning that you can write any algorithm on them. There are many query languages in the market that have limited expressive power, though. Make sure you ask many hypothetical questions to see if it can answer them before you lock in.
Update Data in Real-Time and Support Queries Simultaneously
Graph databases can perform real-time updates on big data while supporting queries at the same time. This is a major drawback of existing big data management systems such as Hadoop HDFS since it was designed for data lakes, where sequential scans and appending new data (no random seek) are the characteristics of the intended workload, and it is an architecture design choice to ensure fast scan I/O of an entire file. The assumption there was that any query will touch the majority of a file, while graph databases only touch relevant data, so a sequential scan is not an optimization assumption.
Flexible Online Schema Environment
Graph databases offer a flexible online schema evolvement while serving your query. You can constantly add and drop new vertex or edge types or their attributes to extend or shrink your data model. Most of the existing queries are still working! It’s so convenient to manage explosive and constantly changing object types. The relational database just cannot easily adapt to this requirement, which is commonplace in the modern data management era.
Make Powerful Recursive Path Query Easily Accessible
Many useful, real-life queries are finding direct and indirect connections in a graph (or network of data). Answering this class of reachability queries is one of the core powers of the graph database. E.g., given a company, find who directly or indirectly invests in the company. Another example, given a product, finding any subparts that are directly or indirectly related to the product.
Further, we can extend the single-pair-vertex reachability queries with multiple reachability queries sharing some common vertices. A set of reachability path queries can be bundled together to constrain each other to form an interesting subgraph pattern. This is the so-called conjunctive graph query (CQ). CQ allows users to come up with a subgraph pattern and asks the database to return all subgraph instances that match this pattern. E.g., given a company, find investors who directly or indirectly invest in the company; and the investors have direct or indirect connection with the founders in the company.
Reachability queries are notoriously hard to do in a relational database, as there is no pre-determined number of JOINs. It becomes harder if we rank the connections (paths) based on some measurement(s) of the paths. E.g., find the shorted path from all flight schedules between two cities; find the person that has the shortest distance to me on the social graph that can connect me to some target user etc. It is easy to use a regular expression to express this class of recursive path queries in the edge pattern of a graph query language. In contrast, it’s very hard to do this in SQL even with the SQL 99 recursive clause (check example here).
Below, we give some examples on a recursive query in GSQL — a graph query language designed for SQL users.
– Finding all investors (firms or individuals) who directly or indirectly invested in a given company within 3-hops. In GSQL, this can be expressed in one line. “Invested_by” is the edge type connecting company and its investors. “<” indicates the source is on the right-hand side of the edge. “1..3” means the recursive range of repeating the Invested_by edge from 1 up to 3 times.
Investors = SELECT tgt
FROM Company:src - (<Invested_by*1..3) - :tgt
WHERE src.name = "TigerGraph";
– Finding all investors (companies or individuals) who directly or indirectly invested in a given company without any upper limits. In GSQL, this can be expressed in one line by removing the upper bound of the repeating edge pattern.
Investors = SELECT tgt
FROM Company:src - (<Invested_by*1..) - :tgt
WHERE src.name = "TigerGraph";
– Finding all investors (individuals) who directly or indirectly invested in a given company, and also directly or indirectly knows the founder of the company. In GSQL, this can be expressed as follows.
Investors = SELECT tgt
FROM Company:src - (<Invested_by*1..) - Person:tgt,
:tgt - (Knows>*1..)- Person:founder-(Founded>)- :src
WHERE src.name = "TigerGraph"
Group by Aggregate Queries
Graph databases, in addition to traditional group-by queries, can do certain classes of group by aggregate queries that are unimaginable or impractical in relational databases. Due to the tabular model restriction, aggregate queries on a relational database are greatly constrained by how data is grouped together. In contrast, graph models are more flexible for grouping and aggregating relevant data. See this article on the latest expressive power of aggregation for graph traversal using accumulators (runtime attributes of vertices and edges, or global states of a query). I don’t think relational databases can do this kind of flexible aggregation on selective data points. (Disclaimer: I have worked on commercial relational database kernels for a decade; Oracle, MS SQL Server, Apache popular open-source platforms, etc.)
Combine and Hierarchize Multiple Dimensions
Graph databases can combine multiple dimensions to manage big data, including time series, demographic, geo-dimensions, etc. with a hierarchy of granularity on different dimensions. Think about an application in which we want to segment a group of a population based on both time and geo dimensions. With a carefully designed graph schema, data scientists and business analysts can conduct virtually any analytical query on a graph database. This capability traditionally is only accessible to low-level programming languages such as C++ and Java. A nice series of webinar make this point clearer.
Graph databases serve as great AI infrastructure due to well-structured relational information between entities, which allows one to further infer indirect facts and knowledge. Machine learning experts love them. They provide rich information and convenient data accessibility that other data models can hardly satisfy. For example, the Google Expander team has used it for smart messaging technology. The knowledge graph was created by Google to understand humans better, and many more advances are being made on knowledge inference. And recently, a GNN concept has been further advocated by DeepMind https://arxiv.org/abs/1806.01261 which emphasizes graph structure data as an optimization configuration of neural networks. The keys to a successful graph database to serve as a real-time AI data infrastructure are:
– Support for real-time updates as fresh data streams in
– A highly expressive and user-friendly declarative query language to give full control to data scientists
– Support for deep-link traversal (>3 hops) in real-time (sub-second), just like human neurons sending information over a neural network; deep and efficient
– Scale out and scale up to manage big graphs
In conclusion, we see many advantages of native graph databases managing big data that cannot be worked around by traditional relational databases. However, as any new technology is replacing old technology, there are still obstacles in adopting graph databases. One is that there are fewer qualified developers in the job market than the SQL developers. Another is the non-standardization of the graph database query language. There’s been a lot of marketing hype and incomplete offerings that have led to subpar performance and subpar usability, which slows down graph model adoption in the needed enterprises. If you want to know further about graph database, download this free ebook which compares many major graph databases’ pros and cons.