THE CLIENT This client has over 55 years of experience as a leader in core banking processing. Their enterprise banking solutions empower institutions to grow their assets and reach new customers by integrating the latest digital technologies. The client's cloud-based core banking platform delivers a seamless mix of innovative, flexible, and secure banking services that provides institutions with a significant competitive advantage. THE NEED: An Agile Payment Solution The client has over 20 years in ATM processing that was founded on two different payments systems prior to the adoption of TANGO. Significantly, the client became increasingly frustrated by both legacy platforms due to the constraints they placed on the client’s business growth. The slowness of vendor maintenance support, the lack of development agility for new customer services, and skyrocketing costs of ownership all led to the need for a better solution for their future. Tired of navigating product and service obstacles with their original payment solution vendor the client turned to a second provider for help. This second provider was able to migrate them to their solution and provided adequate service and functionality for a few years. However, as customer demand for new services increased this second solution had also become stagnant and restrictive. The client then sought a third and final payment solution, one that demonstrably addressed the past growth issues. It was critical that the new solution would provide an agile architecture with the robustness to handle diverse and high-growth volumes. Additional high-priority requirements included; • the agility to continually, and rapidly, adapt to changing consumer needs, • the ability to rapidly create their own configurations on the fly, • a major reduction in application lifecycle costs. THE SOLUTION: TANGO by Lusis Payments The client performed an extensive research of all leading payments solution providers. Their findings indicated that there were decisive advantages in selecting TANGO from Lusis Payments. TANGO and Lusis Payments’ reputation for service and innovation far exceeded all of their business and technical requirements. TANGO's architecture, flexibility, ease of use and lower cost of ownership were once again identified by another client as compelling advantages over other solutions.
In 2012, Lusis Payments conducted a historic proof of concept with partner HPE at the HP ATC (advanced technical center) in Palo Alto, CA. TANGO was tested for 48 hours straight at full capacity. The system processed 2,500 TPS without fail. The hardware configuration used at the time of the benchmark was chosen to match a client’s production system and consisted of a 8-processor HPE J-series NonStop. TANGO proved responsive and surpassed normal daily tasks and nightly settlements. This proof of concept proved that TANGO was fault tolerant and achieved maximum volumes and throughput of a total daily volume of 50 million transactions per day.
The outstanding results came from long hard-working sessions with the HPE teams which we were proud to work with. The first week included our CTO working on-site. Soon after, he was joined by our senior project leaders, and they received significant additional support from our lab in Paris. In addition to the dedicated HPE team, the client’s team also partnered with us to test the conformity of the benchmark protocol. HPE worked with the client to reproduce its environment for a true simulation. It was great project, and we were proud of the outcome. Since then, HPE has continued to suggest that we test TANGO on the newest (Intel based chip) hardware. As we were still quite pleased with the 2500TPS results and the fact that the client continued to realize improved performance on their HPE NonStop platform, we chose not to do additional test campaigns in subsequent years. Until now. At the end of Q1 we said “ok, let’s do it”! At that time, bandwidth was quite low, so we made it “our dry way of doing it.” We used our Vanilla switch based on the TANGO version 7 platform installed on HPE NS server: 8 processor, 6 core NS7-X3 system again at the HP ATC labs. This system runs OS release L21.06.17.2 with NonStop SQL/MX 3.7.2. Each NonStop processor contained 256 GB of memory. We used a very similar testing protocol without any specific tunning. And “Torpedo… LOS”! On our first run we achieved 3,500 TPS. Then with less than 10 hours of tuning, we easily reached 4,500 TPS sustained for two straight hours. So, this has become our new reference on HPE NonStop: 4,500 transactions per second on an 8-CPU machine. And this was simply done with our Vanilla switch and some very light tuning. So, nothing heroic, just the standard product using a standard configuration. In Q3, we will test TANGO with the new Posix Kernel of HPE NonStop and see where we take it! Stay tuned. Philippe Préval Lusis CEO Chez Lusis As more and more companies choose TANGO to replace their aging legacy payments systems, we thought we'd shed some extra light on it by talking with Brian Miller, our General Manager at Lusis Payments. The consumer demand for faster, more secure 24/7 payments continues to challenge payments organizations across the globe. For many organizations, the costs and difficulties of nursing an aging payments platform are now unsustainable. As a result, these organizations are now planning to replace their legacy applications with a faster, more agile solution that can free their business from the constraints of an inflexible payments platform. Lusis Payments has set out to help payments organizations simplify and streamline their migration projects, reducing risks and delivering predictable progress milestones. As the innovative provider of TANGO, the mission-critical online transaction processing engine, Lusis Payments is making it much easier for organizations to keep pace with consumer desires for greater convenience, speed, and security in payments. Q. What are the biggest challenges organizations with legacy transaction processing systems face today? Miller: Legacy systems bring a whole host of challenges for organizations. For example, one of our clients, a top-five global bank, found their 28-year-old legacy system severely constraining to their business. It had become too expensive to maintain and operate, development times were lengthy, and they were less able to compete in the market. When they started looking to replace the system, they realized that they were using the software throughout their entire line of banking services. Clearly, it would be crucial for the new solution to provide a highly extensible architecture, enable them to orchestrate low-risk migrations, and be powerful enough to handle diverse transactions and increasing volumes. Of course, they also wanted a solution that would reduce their cost of ownership and application life-cycle costs while increasing the bank's agility in adapting to consumers' changing needs. The ability to efficiently support new regulations and scheme mandates was another key requirement. The bank conducted an extensive analysis of the leading payment solution providers. The analysis showed that TANGO exceeded all the client's business and technical requirements, and TANGO outperformed its competitors in the areas of architecture, flexibility, and cost of ownership. Q. Why is TANGO so successful in replacing legacy systems, such as BASE24®? Miller: Much of Lusis Payments' success comes because TANGO is easily “built to order” because of its micro-services platform. We recognize that payments organizations need a solution that works the way they do, that empowers its staff, not hinder them. TANGO does this. A good example of this in action is our client BankservAfrica, which wanted to expand into the rapidly growing South African development community. Well, to do this, they needed a fully functional core system that could cope with fast-changing payment methods and customer requirements. TANGO also met BankservAfrica's business requirements, which included that it must be configurable, with specific monitoring capabilities. In addition, TANGO's cost and clear licensing structure appealed to the BankservAfrica team and our phased approach really makes for a painless migration. Q. It really says something that some of the largest banks in the world have chosen TANGO to replace their systems. What's something you want other organizations and financial institutions to know about TANGO as they may be looking at replacing legacy systems? Miller: At Lusis, it boils down to this. We don't care what transaction processing system you had. We want to hear about what functionalities you want for the present and the future. With TANGO, we can build whatever you need with scalability for any of your future needs as well. Q. What should organizations avoid when it comes to migrating their payments system? Miller: The most important advice I can give is don't wait. To be successful, organizations need to develop the operational skills and procedures to manage continual change. Migrating to a new payments solution is not a one time thing, it is actually a transition to a different operational lifestyle – one where change is the expected norm.
THE CLIENT This Lusis Payments customer is one of the largest financial institutions in the world. The organization has over 10 million global customers covering several public and business sectors. They provide tailored banking solutions for personal, small business, commercial and cross border payments. THE NEED: Freedom from a Constraining Legacy Payments Platform The client's business was becoming severely constrained by their 28 year-old legacy payments system. The high cost of ownership, lengthy development times, and soaring maintenance challenges became urgent pressures for change. The legacy software was widely utilized throughout the client's vast line of banking services. It was therefore crucial that the new solution would provide a highly extensible architecture, facilitate low-risk migration projects, and have the robustness to handle diverse and high-growth volumes. Additional high-priority requirements included;
“We have implemented more “new” functional capabilities in the last 3 years on the TANGO platform than we have in over a decade on our previous legacy platform. It is great to be free from all the constraints.”
By Fabrice Daniel, Artificial Intelligence Department of Lusis, Paris, France http://www.lusisai.com ABSTRACT
Combining evidence from different sources can be achieved with Bayesian or Dempster-Shafer methods. The first re-quires an estimate of the priors and likelihoods while the second only needs an estimate of the posterior probabilities and enables reasoning with uncertain information due to imprecision of the sources and with the degree of conflict between them. This paper describes the two methods and how they can be applied to the estimation of a global score in the context of fraud detection. INTRODUCTION Fraud detection mainly relies on expert driven methods that implement a set of rules and data driven approaches implementing machine learning (ML) models. Both provide an estimate (or a score) for a new transaction to be fraudulent. While each ML model naturally returns a fraud probability, the experts can also attach a probability to each rule. They can also be automatically calculated from the labeled his-tory. Combining them together produces a global score that can be used in a near real time system to rank a set of trans-actions having the highest probability to be fraudulent. By obtaining this ranking, investigators can concentrate their efforts on the suspect transactions with the highest probability of being true frauds. The most common approaches for combining scores are summing individual scores or returning the highest score among the trigged rules. This is not entirely satisfactory given that summing scores is equivalent to averaging the probabilities returned by each predictor (rule or model). It also does not take into account the uncertainty of each predictor and the degree of conflict between them. For the Lusis fraud system, we work on implementing more appropriate approaches. This paper proposes two ways for addressing this problem. The first is to use Bayesian methods [5]; the second is to combine the scores by using Dempster-Shafer theory [6]. |
lUSIS nEWSThe latest company and industry news from Lusis Payments. Archives
March 2024
Categories
All
|