Accelerating HPD Lendscape Factoring Services with The Hazelcast Platform

How a FinTech company leveraged in-memory technology to speed up its software offering.


Software vendor

Year Founded



Hazelcast Platform

Factoring is a financing model whereby a finance company purchases another company’s outstanding invoices to provide funds quickly without having to collect the debt. Acquiring loans can be onerous and take a long time to set up. Without factoring, small businesses wouldn’t be able to survive.

The factoring market size value is approximately $3,235.88 billion (USD) globally, as of 2020 with a projected growth to $5,384.00 billion (USD) by 2027.


HPD Lendscape is one of the world’s leading software vendors for secured business-to-business lending software used by over 140 banks and financial institutions across the world. HPD Lendscape’s platform, Lendscape, offers a range of capital financial solutions such as factoring, invoice discounting, supply chain, and asset-based lending.

One of their key clients is Lloyds Bank Plc. The original solution was developed in Cobol running on a mainframe platform but has since been re-developed in Java to modernize the platform and make it more appealing to customers focused on digital transformation. Lendscape is deployed on-premises and will soon be available in the public cloud. For hosted solutions, HPD Lendscape rents data center storage across the world and they own the hardware in the center.

The platform is managed across three core teams, engineering, technical services, and the internal IT department which consists of almost 100 experienced developers, support, and operations personnel. The engineering team is run by an experienced leader and consists of 75 individuals, broken into squads who own functional areas of the system. Each squad is comprised of a mix of disciplines, i.e., analysts, testers, and developers, who are semi-autonomous. They can be deployed to address issues very quickly following a typical agile model. This gives enormous flexibility, and each squad can work on any of HPD Lendscape’s established solutions or upcoming projects.

As a very agile business, HPD Lendscape is always engaged in a cycle of continuous improvement. They are constantly seeking better ways to do things and are not afraid to fail in the process.


It was clear from the start that the development of HPD Lendscape would need a solution that was able to continuously process high volumes of data as fast as possible. The solution also needed to be scalable so that it could accommodate all sizes of factoring solutions. Reliability was of course a critical factor and having downtime was not an option. To support their agile teams, the solution needed to be easy to understand, build, and deploy. In addition, the application has evolved to use the Spring Framework as well as other widely adopted frameworks in order to better support the future maintainability of the application. As the market has evolved and adapted to changing technologies, so too has HPD Lendscape’s solution. The most recent evolution has focused on cloud adoption and to this end, HPD Lendscape is able to host on AWS and Azure.

Initially, Ehcache was used as the in-memory solution deployed for Lendscape and it has continued to prove sufficient for single-server deployment. However, to provide solutions for the future that allows for scalable multi-node deployment, they needed a caching solution that would support real-time transactions, with very high throughput and low latency and would run continuously without service outages.


Hazelcast was one of several caching solutions that were being reviewed against strict criteria that included:

  • Ease of use – quick deployment
  • Familiarity of toolset
  • High performance, low latency
  • Cloud readiness
  • Resilience

After benchmarking several solutions, Hazelcast was chosen. Hazelcast’s integration with the Spring Framework and Azul Zulu OpenJDK helped accelerate the deployment of Hazelcast.

Hazelcast is embedded within every installation of HPD Lendscape’s solution. Hazelcast’s distributed architecture was a key factor in their decision-making as it offered future deployment and scalability options. Currently, in the core product, there is one central processing engine that handles all processing. The longer-term plan is to split this into multiple engines that are unbounded. In addition, the portability of the platform affords the luxury of being able to deploy Hazelcast in several different ways moving forward, including changing the architecture for their front-end processing.

Hazelcast is deployed within Lendscape’s backend services to accelerate data processing from the main engine database and the reporting databases that are hosted on predominantly MS SQL Server, DB2, and Oracle. The main engine database is multiple terabytes in size and the reporting database is often up to five times bigger and stores a normalized view of the data. When the service starts up, the data is “eager” loaded (i.e., loading the data when the Hazelcast cluster members startup, rather than upon request) to ensure that the in-memory component is fully populated with all static data required to serve the backend and two front end clients. For HPD Lendscape, it is crucial to have a caching solution that holds all the data in the cache and allows capacity increases without compromising on performance—Hazelcast’s distributed architecture supports linear performance with growth.

Hazelcast is primarily used to store static metadata within the system, such as “translatable strings” that are presented to both clients and their customers. These “translatable strings” are literals within the system to accommodate multi-language support (for example, a French client with Dutch customers, or a client in Israel using Hebrew to display the text “right-to-left”) as the software is used all over the world. Within this complex computing system, there is a lot of metadata data being constantly being used within the system.

All the data stored within Hazelcast accelerates data access to serve both the two ancillary front ends; Aura and the Client Manager—both are web clients supporting the internal and external web clients enabling invoice access and other functions, but also the core application

HPD Lendscape currently does not utilize any of Hazelcast’s enterprise features like encryption, WAN replication, persistence, blue/ green deployment, and auto-failover out-of-the-box. However, if a client wants to implement a security model, provide extremely high uptime, deploy to the cloud, or geographically disperse their data for client locality and high availability, they can purchase an enterprise license from Hazelcast under the full support of HPD Lendscape.


The Hazelcast Real Time Stream Processing Platform has ensured that the database access is no longer the bottleneck—the platform is now only limited by its hardware constraints. The feedback that was given to Hazelcast by the Head of Engineering was, “The good thing about Hazelcast is it just works…” HPD Lendscape has seen a much-reduced number of issues around data access speeds since deploying Hazelcast enabling the system to process data well within SLAs. As a real-time, transaction processing system, the need for performance is pervasive. Having the best performance can differentiate your business from the competition and is critical to the success of the business. Deploying Hazelcast has offered other benefits–they were able to benchmark against the European volume of data for factoring using data generation tools that simulate real-time processing and conduct several tests, including soak and performance tests, uploading invoices and debtors, for example.

The predictions for factoring suggest that there will be an even greater demand to reduce the elapsed time from application to cash-in-hand. In addition, as the uncertainty of the economic environment continues, more online onboarding and direct digital payments are predicted to be essential for providing capital quickly. As a result, factoring companies will need to ensure that they can manage risk effectively.

The latest version of Hazelcast offers both in-memory storage for fast access as well as stream processing capabilities as one solution with one codebase. Combined, these capabilities offer critical use cases such as fraud detection and risk mitigation. Hazelcast’s machine learning capabilities could also play a part in reducing risk and minimizing fraud. In addition, Hazelcast will be extending its High-Density Memory Store capabilities so that even more data can be stored per server, thus reducing the overall footprint.

As Lendscape evolves with improvements and new features, supported architectures like the Hazelcast digital integration hub will enable commonly used data to be shared across several different services to simplify integration with other applications, including open accounting, giving the client the ability to access ledger information from online accounting packages like Sage and Xero installed on end-user PCs and then uploaded to the cloud to drive funding decisions.

Overall, their strategy is to move to the cloud and continue to provide fast, secure, and user-friendly solutions at the speed of the market. However, one factor that must be considered is “data sovereignty”—abiding by the laws in the country where that data is processed is critical. Either way, HPD Lendscape’s journey with the Hazelcast platform has the potential to grow in volume and scope, helping to keep Lendscape at the front of the factoring market.

The good thing about Hazelcast is it just works.

— Head of Engineering, HPD Lendscape