Re-architecting your platform for a competitive edge: A strategic approach

By Michael Babushkin, the Founder and CEO of Devexperts

A new era of trading platforms is emerging

In recent years, end users’ behavior towards electronic trading platforms has drastically changed due to the overall developments in our everyday life. The 100% adoption of smartphones, companies embracing remote working during and after the pandemic, the disruptive changes in the retail brokerage landscape with zero-commission online brokers gaining popularity, and finally, the newly available technologies knocking on our doors at any given moment (blockchain, AI, AR/VR, and metaverses) have all impacted clients’ perceptions of online trading platforms.

Therefore, typical users no longer expect legacy desktop-based feature-rich trading platforms, but instead, an ecosystem of services and data summaries available at their fingertips. Still, the newest version of trading platforms requires market analysis tools to develop their trading strategies, as to not leverage the advice of a third party.

These new technologies such as cloud-based computing, data lakes and data oceans, modern powerful mobile devices, and AI, improve or completely change how firms can optimize market analysis.

AI itself can be the next technological revolution, and ignoring its potential would be a mistake for the electronic trading industry. Having said that, Devexperts encourages brokers to adapt their current solutions to the evolving technologies and to the changing end-user behavior. Can it be done gradually? We believe yes, and this is what we call re-architecting.

Although there have been no obvious breakthroughs in IT over the past few years, there are major trends in AI and cloud computing that have been observed that can give trading software systems a competitive edge.

Knowing when to revamp

Although there may not be a ‘specific’ moment in time where a platform’s revamp is encouraged, there are clear signs that a system is due for a redesign, re-architecture, or refactoring (even if the system fulfils its business requirements). Below are some examples of circumstances that signify a change is needed:

• Significant performance issues during operation.
• Implementing new requirements takes too much effort.
• Maintenance and support do not work smoothly.

Developing the optimal technology stack

If we were to create the ultimate technology stack, with no limitations to budget or skill, we would certainly recommend using Java. With our expertise in the Java ecosystem, it is safe to say that Java is great for developing business applications due to its stability, huge variety of frameworks and libraries, and integrations with third-party systems.

However, Kotlin by JetBrains may be an even greater alternative to Java, as it is 100% interoperable with Java (and therefore it inherits the entire Java ecosystem). Kotlin promises to become more than an alternative, but something bigger.

Its multiplatform capabilities, compiler and interoperability layers already allow the creation of web and mobile applications using Kotlin programming language only.

Finally, Kotlin simply performs better than Java – it is modern, concise, and functional as well as object-oriented. The more homogeneous and standardized the technology stack is, the easier to recruit an engineering team and build the development processes. Also, the better it is for the interchangeability of team members.

A more realistic approach

Now that we had a look at the most optimal technology stack, let’s get back to reality and focus on a more realistic and manageable trading platform re-architecting method. Below are some best practices firms can implement for a successful revamp of their platform.

1. Reusing existing code: To save costs, it is important to reuse as much of the already existing business code as possible. Instead of copying/pasting, the code should be optimized in terms of efficiency for using algorithms and data structures, as well as resource consumption. It is important to note that your code must be well covered with tests for the refactoring process for it to smoothly be implemented into the re-architecting. Additionally, developers should avoid re-inventing the wheel, instead, use industry-standard protocols, best practices, and approaches for typical tasks to save time and resources.
2. Embracing microservices: Split the monolithic system into smaller components for better microservices. The microservice approach allows developers to scale individual components, achieving better performance while saving computing resources.
3. Utilizing cloud technologies: Take advantage of cloud technologies, containerize your applications, and equip them with distributed configuration, standardized logging and performance metrics to allow deployment and hassle-free operation on cloud platforms. Well-established CI/CD infrastructure (Continuous Integration/Deployment) is time-saving and it is overall much easier to maintain development, test, and QA environments in a private cloud.
4. Recognizing fully extensible architecture, as an ideal: A best practice would be to only have one core of the system, which defines the standard. All other components (e.g., customer-specific functionality, paid add-ons) are developed as extensions of that standard. Although extensible architectures may take more time to develop, the approach is very beneficial since product knowledge is better structured, and customer-specific development teams do not have to understand all the system code, they only need to know how to use it through extension points.

To build or buy?… or both!

Although there may be some benefits to building a technology stack from scratch, due to the lifetime of modern IT solutions getting shorter and shorter, it is important to consider that by the time the stack is built from scratch, it may very well already be outdated once completed. Therefore, buy or hybrid methods are recommended.

Now that isn’t to say that building is a bad thing, however, it is simply not worthwhile for a broker to build themself, especially if developing a large upgrade that may require years and substantial monetary investment. Buying or going hybrid often leads to a quicker go-to-market implementation.

On the other hand, opting for the buy method doesn’t mean you must stick with a boxed solution. Devexperts works with a number of brokers who leverage our products as the foundation for their solution, then build on top either with a vendor or with a hybrid model with their teams jointly working.

We’ve seen great success in both models, and it ultimately comes down to the ownership and investment a firm wants to make for its platform. If an extensible architecture approach is adopted, the optimal structure would be that the core team is in-house, while extensions (given the standard is well documented and supported) can be outsourced.

What are the 3 factors of a successful platform re-architecture?

Interoperability

To Devexperts, interoperability heavily focuses on the connectivity of the ecosystem itself. It is crucial to ensure all aspects are working together seamlessly (API).

To ensure interoperability, software architecture should define the following key aspects:

• Data Interoperability: Establishing unified structures for exchange data
• Functional Interoperability: Enabling components to efficiently work together while performing a task
• Protocol Interoperability: Clarifying how components communicate over a network, architectural documents define protocols used in both the system and API contracts
• Platform Interoperability: Facilitating operations across diverse platforms and environments such as desktop, web and mobile.

Resilience

To ensure your platform’s revamp is successful, considering high availability and Disaster Recovery (DR) scenarios while developing and planning deployment is an important factor for optimal resilience. In conjunction with development, estimating and foreseeing the redundancy of current components and environments is another sure-fire way for a trading platform to maintain resiliency and stay ahead of the curve.

Profitability

Maximizing profitability while optimizing resources remains a top priority for businesses, particularly in the realm of software development. Devexperts firmly believes that a solid foundational architecture is the key to achieving both goals simultaneously, driving maximal profitability through resource optimization.

The right architecture makes a product easier to support and maintain. Hence it reduces time and costs associated with operational activities, leading to faster time-to-market. Furthermore, an architecture designed with extensibility in mind facilitates the implementation of new features or modifications with fewer resources.

Leveraging industry-standard components and best practices not only streamlines engineering staff recruitment and training but also makes teams more interchangeable. Additionally, software architecture may affect the ease of integration with other systems, directly correlating with potential expansion. The easier the integration the more business opportunities.

Overall, good architecture results in a more reliable and secure product, with less downtime and better user experience, which can increase customer retention and attract more customers. Security is key to maintaining customer trust and avoiding potential losses associated with breaches or theft of data.

What tools can be utilized to achieve interoperability, resilience and increased profitability?

We’re happy to share our technology stack, but we’ve seen successful adoption of alternative stacks, showing as great usability and performance as Devexperts solutions. Examples of some of the many proven tech stacks are:

• Programming languages: Java, Kotlin
• Front-end / multiplatform frameworks: ReactJS, Kotlin
• Run-time deployment platforms: Kubernetes
• Messaging systems: Quote Distribution (QD) binary protocol (open source, created by Devexperts) for stateless data transmission; Kafka for stateful data transmission.

What are the business and operational benefits of getting the technology stack right and the penalties of getting it wrong?

Benefits of getting it right:

• Cost reduction: Reduces development cost with the extensibility of the core system
• Ease of scalability: Easier to scale up with the correct technology already put in place
• Forming the right team: Easier to find personnel to work on the systems, whereas relying on antiquated resources tends to be more expensive and challenging.

Penalties for getting it wrong:

• Reputation: outages may lead to loss of confidence/trust in the industry, in turn, leading to loss of customers
• Financial penalties: May need to pay fines if the firm is unable to keep sensitive data safe
• Obsolescence and lack of support: Utilizing outdated or unsupported technology may cause the platform to become obsolete as more advanced technologies emerge
• Constricted potential for growth and innovation: An incorrect technology stack may struggle to keep up with the evolving industry. This can restrict the trading platform’s capacity for growth, success and opportunities for innovation.

When choosing a technology stack, it is important to keep in mind: 

• Availability of technical specialists working with the stack
• Number of successful adaptations of the stack in your industry and overall
• Plans for improvements/scaling of the current solution and budget for changes
• Uptime and security requirements, depending on the application. The higher they are, the more expensive the development can be for a particular technological stack.

Key Takeaways

When developing a trading platform, it is important to think about modularity. It is beneficial for firms to build components of the solution as standalone applications, which can then work together to create the ultimate technology stack.

Developers don’t need to reinvent the wheel. Stick to industry standard frameworks and infrastructure components to save time and resources. Respect those who will support and maintain your product is crucial, by making the product observable and transparent in terms of logging, performance metrics, and configuration and by constantly improving the code to avoid the stack becoming obsolete.