A Temenos white paper
Everyone’s
Banking
Platform
Temenos High-Water
Benchmark on Temenos
Banking Cloud
Temenos Product Performance and Benchmarking Report
August 2022
Contents
1. Abstract 3
2. Introduction 4
3. Architecture 6
4. Temenos Banking Capabilities 7
5. Summary 10
Annex: Scalability, Benchmark results and Metrics 11
6.1 Transaction Linear Scalability 12
6.2 Benchmark Result 13
6.3 Benchmark Metrics 14
2
The target to the Highwater benchmark was to achieve
100,000 Transactions per second with Decoupled Temenos
Banking Capabilities on Temenos Banking cloud. The key to
Abstract
the benchmark was to use serverless technologies along with
the 2 years of platform innovations in Temenos.
Throughout the document, we will use few acronyms
which describes the Product components, Monitoring and
Benchmarking terminologies which are illustrated below.
TBC Temenos Banking Capabilities
TPS Transactions per Second
API Application programming Interface
DDA Demand Deposit Accounts
3
Introduction
In May 2022, Temenos published the high-water benchmark performance results for Temenos Banking
Cloud on Amazon Web Services (AWS). The benchmark encapsulates two years of additional innovation
at Temenos since the previous benchmark in 2019, resulting in significant advancements for a leaner and
greener architecture.
This benchmark breaks every scale record, helping banks scale efficiently and lower carbon emissions.
The results are a testament to this: 100 million customers and 200 million deposit accounts with 100,000
transactions per second and up to 61 transactions per second per core.
This improvement in TPS per core also shows up to 4.1x times more efficient compared to the
already sweeping benchmark results of 2019. The additional two years of innovation return a doubled
performance, proving the scalable performance of the Temenos open platform with the capacity to
process the entire world’s banking transaction volumes on a single cloud environment.
In this report, Temenos provides more data and insights to show how the Temenos Banking Cloud
architecture elastically scales, automatically enabling banks to process higher volumes of transactions
with 10x less infrastructure than traditional on-premises deployments and a corresponding reduction in
cost and carbon footprint.
4
Benchmark stack: Temenos Banking Cloud, AWS, About Temenos
Yugabyte and MongoDB
Temenos supports 3000 banks around the globe, including 41 of the top 50
global banks and over 70 Challenger Banks. Temenos has been pioneering in
The benchmark was performed on the Temenos Banking Cloud with Decoupled
the cloud since 2011 and was the first software provider to put a core banking
Temenos Banking Capabilities on cloud native composable architecture and
system in the public cloud. Today Temenos supports over 700 SaaS clients and
serverless technologies.
dedicated $1.0B for R&D in the coming 3 years.
The benchmark proves that Temenos Banking Cloud running on AWS and using
database technology from MongoDB and Yugabyte enables banks to harness the
power of the cloud and benefit from speed, scalability, resilience and innovation.
First class business events were generated with Amazon Kinesis as the
streaming platform. Redis was used as a managed service for external caching.
The benchmark tested Temenos composable Banking Capabilities including
Demand Deposit Accounts and Loans. Temenos Banking Services include BNPL,
Challenger Bank, Retail Lending, Payment Fraud Mitigation, KYC Customer Risk
Assessment, Digital Mortgages and Demand Deposits Accounts.
5
Architecture
• Amazon Web Services platform with Amazon Kinesis streaming platform and Redis ElastiCache
• DDA, Reference Data, Market Data TBCs in EKS cluster
• Holdings, Party TBCs as Lambda functions
• YugabyteDB as DBaaS
• MongoDB as DBaaS
6
Temenos
Banking
Capabilities
7
REFERENCE DATA TBC PARTY TBC
Central system to store and service Reference data used by multiple business Party TBC provides an independent cloud-enabled solution for managing
products. It stores common lookup data that are non-volatile in nature like party information that is required for both front office systems and back office
Country, Currency, Region, Holiday that are refered in Temenos services. systems.
Reference Data TBC also includes the data published periodically by Regulators,
Markets and Payment providers like BIC, LEI, IBAN, Bank codes. Party TBC holds the personal information related to the Customer like the
Rest APIs are used to maintain and service the data in Reference data TBC. addresses, contact information, employer information, Business information, Tax
Regulation Details, and many more related data on a real-time to the end-user.
YugabyteDB was used for Reference Data TBC in this benchmark.
Party TBC was deployed as an independent serverless TBC in this benchmark.
MARKET DATA TBC MongoDB was used for Party TBC in this benchmark.
Central system to store and service Market data used by multiple business
products. It stores volatile data published by Market like currency exchange
rates, interest rates, future exchange rates that are refered in Temenos services.
Rest APIs are used to maintain and service the data in Market Data TBC.
YugabyteDB was used for Market Data TBC in this benchmark.
8
HOLDINGS TBC EVENT STORE
Holdings TBC provides an independent cloud-enabled solution for managing account information that Event store provides Event-driven architecture across
is required for front office systems. Temenos TBCs. The events raised are recorded as part
Holdings TBC holds account information like Balance details, Transaction details, Portfolio details, of Inbox events and raised it to respective TBC using the
Allocations of an account. outbox infrastructure.
Holdings TBC was deployed as an independent serverless TBC in this benchmark.
YugabyteDB was used for Event store in this benchmark.
MongoDB was used for Holdings TBC in this benchmark.
DEPOSITS TBC (DDA)
The Accounts Temenos Banking Service is a unique composable banking service that enables financial
institutions of any scale to create, manage and service savings and current/checking accounts.
Complete with the ability to integrate with a complex manufacturing and distribution landscape, the
Service covers the quick and flexible definition of products, as well as processing of such products
across their lifecycle.
The service supports numerous transactional and servicing capabilities including interest and fee/
charge processing, as well as tax. It is based on a headless architecture with a published set of APIs
and events for account opening and the various life cycle/service operations of the account.
YugabyteDB was used for Deposits TBC in this benchmark. 9
Summary
For all details on the performance, we refer to the Annex:
Scalability, Benchmark results and Metrics
The benchmark of 100 million Customers and 200 million Accounts with 102,875
transactions per second and up to 61 transactions per second per core on AWS
underscores Temenos’ leaner and greener architecture.
22% lesser hardware used when compared to 2019 High-Water Benchmark with double
the throughput.
Funds Reservation response time at 65 ms and throughput at 61 TPS per
core which is 4.1 times faster than 2019.
Funds booking response time at 136 ms and throughput at 35 TPS per core
which 2.8 times faster than 2019.
Query response times at 16 ms demonstrating the high performance of
Temenos Banking Cloud.
10
Annex: Scalability,
Benchmark results
and Metrics.
11
6.1 Transaction Linear Scalability
Below graphs show the linear scalability of the transactions where the X axis
denotes the number of pods and Y denotes the TPS
12
6.2 Benchmark Result
Below benchmark result show an overall TPS of 102,875 where the peak TPS was
reached in 30 mins and a steady state of peak TPS 102,875 was maintained for 30
mins. Refer the below table for the TPS of each transaction in this benchmark.
Transaction TPS Response Mix
time (in ms)
Payments 465 686 1%
Reservation 8,750 65 9%
Booking 12,760 136 12%
Transfer 1,485 248 1%
Balance Enquiry 47,315 15 46%
Transaction List 32,100 17 31%
102,875
13
6.3 Benchmark Metrics - application
Below shows steady Response times (Latency) for the transactions with healthy CPU utilization and memory
utilization for the Application layer
Application
14
6.3 Benchmark Metrics - Yugabyte
Yugabyte stats show Op latency under 4 ms with steady CPU, Network and memory utilization
yugabyte
15
6.3 Benchmark Metrics - mongodb
MongoDB stats show the disk latency is under 2 ms with steady CPU and memory utilization
mongodb
16
