An eCommerce Order Management System (OMS) is a software application designed to manage the entire order process, from receiving and processing orders to fulfillment and delivery. OMS is an essential component of any eCommerce business that requires order tracking, inventory management, and shipping management
OMS automates the entire cycle of business processes associated with creating, processing, and executing orders. This article will discuss the peculiarities of OMS architecture for eCommerce stores with over 500,000 orders per day especially when it comes to inventory software management.
Table of Contents
What is eCommerce OMS from a technical point of view?
OMS is an event-oriented processing system. The processing unit is a lot, an indivisible trading unit. Each item the customer adds to the order is allocated in the OMS as an independent lot in terms of processing connected to the order within which the lot exists.
Every OMS requires process mapping. It describes the lifecycle of every lot and is essential for handling orders. It reflects all business processes, their branching, dependencies, and reactions to events. These facts took place in a physical or virtual world, for example, delivery to the customer, order cancellation, etc. It’s also vital to take care of your computer for it to work fast and help you with all your tasks. So, if you have system junk, unwanted apps, etc, you can check Clean My Mac X.
Process Mapping: What does it consist of?
Each process contains a predetermined set of possible steps (nodes) that the lot can go through. Every OMS uses its internal notation to describe the nodes of the process map. They come in several types:
- It is a node for performing specific, predefined actions: an integration call, DBMS call, etc.
- It is a node of process termination.
- A node implements the suspension of lot processing until a certain absolute or relative time point is reached.
- It is a node without a functional action, describing the beginning of the process map section responsible for processing some occurred event.
- It is a node without a functional action, providing a hierarchical grouping of other nodes.
So, process mapping is one of the most important parts when building order management software.
OMS load: approximate calculations and a bit about processing assurance
Imagine that there are 1 million daily orders with 2 products in each. As we already know, the processing unit is a lot, so our orders will have 2 million lots. Let’s assume that each of them will pass 1/3 of the available map processes (1,500 processes out of 4,500). Thus, the number of actions performed per day would be 3 billion.
Assuming that each action is just 1 simple integration operation or DBMS access, and the load distribution is strictly even throughout the day, we get 34,722 operations per second. And that’s without taking into account API traffic, which OMS provides.
OMS Infrastructure: how to get better performance
Physical hosts and virtualization are often used to deploy services. With the high load that OMS generates, deploying, for example, SQL on virtual hosts is not the best option. Technically it is possible, but to end up on the same host with some full-text search engine like ElasticSearch is fraught with unrealistic freezes. OMS will perform best on physical servers with Intel Xeon Gold, SSD, and 100 GB DDR5.
Some eCommerce order management software also uses virtualization. Suppose that among the eCommerce store business models, there are operational and non-operational models:
- Operational models are when the eCommerce store takes care of organizing logistics. Accordingly, many resources are required for processing and interaction with delivery services.
- Non-operational models are when the eCommerce websites do not need to organize logistics. For example, in order pickup format, the buyer reserves goods through the eCommerce store at a convenient location and then picks them up on his own.
So, non-operational models are the first candidates for placement in virtualization because they require fewer resources to process the execution of orders.
OMS DevOps
The quality of OMS development and high time to market is achieved using methodologies for automating the process of testing, configuration, and deployment. OMS deployment consists of 4 parts.
Changes Deployment in the database structure
Most platforms provide restructuring of tables, but sometimes it may be appropriate and manual intervention. For example, it is relevant to change large database tables. Not frequent, but a very real case.
The regular platform restructuring mechanism on large tables is historically slow. Since the operation is performed in a monopoly mode of access to the database, it is fraught with long downtime and inaccessibility of the system. To speed up the process, replacing the original large table with a structurally identical but empty one is popular.
Thus, the technological process of restructuring on the part of the platform is performed almost instantaneously. The large table with the data is changed using manual scripts, within a virtual data room, which is often more optimal and faster without stopping the service.
Code Deployment
Deploying code to a working system without stopping the service (“dynamic update”) is possible. Most developers and platform administrators consider it a harmful procedure, but sometimes it is impossible without it.
Also, one of the successful solutions for removing bugs on a running service is the mechanism of platform extensions. To quickly fix the problem, developers install an extension with a bug fix. And after transferring changes and deploying the main configuration, the extension is removed.
Process Map Deployment
Lots and orders are executed according to the implemented processes. The process map, a dynamic component of the Operational Management and Reporting (OMR) system, holds a pivotal position. As an integral part of the system, the process map, accompanied by streamlined order tracking, plays a crucial role. This inclusive approach ensures that any discrepancies or irregularities are minimized, preventing potential inaccuracies in processing and safeguarding against the loss of orders.
When deploying a process map, it is imperative to establish a robust and efficient order tracking system. Such an approach not only facilitates the seamless processing of upcoming orders and lots but also guarantees the smooth progression of existing orders within the system. By emphasizing the importance of an effective order tracking mechanism, the overall efficiency and accuracy of the system can be significantly enhanced.
Additional Settings Deployment
Some additional settings of the system, affecting the logic of its work, are stored in the database. These settings are integral to the deployment and may depend on the dev/test/prod environment or, for example, on the target cluster operating model. For example, if we are talking about operational and non-operational models of eCommerce store operation.
Conclusion
Custom OMS is a powerful and flexible tool for business support. It is a unique product from a business and technological point of view. The synergistic combination of sometimes not-at-all-obvious stacks, technologies, and approaches creates room for creativity and encourages continuous search and growth.
