Petabyte-Scale Data Management for Cloud Banking
Total Page：16
File Type：pdf, Size：1020Kb
WHITE PAPER PETABYTE-SCALE DATA MANAGEMENT FOR CLOUD BANKING Migrating banking data from legacy systems to the cloud is easier said than done, especially if data volumes are in petabyte scale. A change in people, processes and perception is required before banks can begin their journey to the cloud. Navigating these is crucial to a bank’s transformation and survival in this customer-centric age. Banks view Challenge of cloud cautiously cloud banking Cloud computing has fueled rapid automation, and analytics. According growth and adoption of innovative 1. Legacy technology to Gartner estimates, banks need to digital business models. From Uber to According to an Infosys Knowledge triple their digital business innovation Spotify and Netflix to Slack, thousands Institute survey, legacy systems budgets to modernize legacy of digital leaders have built their currently rank as the third most applications through 2020.4 businesses on cloud software. This is commonly cited barrier to digital because it enables flexible, scalable, transformation (named by 42% Experts say that banks spend 80% of cost-efficient and continuously of respondents), but financial their IT budgets on legacy technology upgradable capabilities. But for banks, services executives expect that they maintenance, and a Tier 1 bank could adopting cloud technologies can be could become the most serious spend up to $300 million a year to a real challenge — not least because barrier in 2019.2 update existing software to meet mainframes continue to be at the core regulatory requirements.5 But cloud Millions are spent every year to of their business, given their reliability, can act as a workaround in reducing maintain these archaic systems,3 which almost impregnable nature and ability costs.6 For example, the Federal are often incompatible with modern to manage large volumes of complex Home Loan Bank of Chicago reduced technology. These systems act as a tasks. In fact, over 90% of the world’s infrastructure costs by 30% by moving barrier to better catering to the needs top banks still rely on mainframes.1 all of its internal production workloads of the digital customer. The reliance This is despite the fact that mainframes to the cloud.7 on legacy technology also stops banks are archaic in nature and present a from taking advantage of Agile and host of issues. DevOps ways of working, increased External Document © 2021 Infosys Limited 2. Data management copies created in the data lake and 6. Agile and DevOps transported to different systems either With data growing by 2,500 ways of working in the cloud or internally. petabytes a day,8 data management is Legacy systems are expensive to increasingly becoming a concern for Banks’ manual and siloed approach maintain and delay a product’s time banks,9 not least because of regulatory of ingesting and migrating data to market.17 They are also not ideally and security requirements. Regulators is time-consuming. It is also suited to Agile programming methods, require banks to provide detailed counterproductive, as in this digital instead relying on waterfall methods reports, additional information on age, customers’ demand for instant, that can slow down the production stress testing and information at real-time, and personalized products of software and result in less timely a granular level.10 Banks continue and services has become the feature releases. to capture data in multiple forms new normal.13 Banks that have made the shift to (customer personal information, Agile and DevOps ways of working transaction history, journey maps, 4. Lower number of have benefited. In 2012, J.P. Morgan market data), and this huge volume of legacy coders followed a quarterly software-release data is now stored in data warehouses Legacy systems are built on outdated cycle, and coordination between and lakes. languages such as COBOL. For development and operations was Banks store data in various locations, instance, the U.S. financial sector minimal. These quarterly releases and it’s often accessed by different has over 200 billion lines of COBOL heightened risks, were cumbersome users, creating a siloed approach and code currently active, and COBOL and time-consuming, and increased multiple layers of data duplication.11 powers over 90% of ATMs.14 Despite delivery costs. The financial institution Singular data warehouses and COBOL’s widespread use, today’s decided to embrace Agile and DevOps lakes are formed, giving rise to coders prefer to use new languages practices. The software-releases further issues of storage, meeting that are compatible with artificial cycle changed from quarterly to 100 regulatory requirements, authenticity, intelligence, machine learning or cloud releases in 2015, 200 in 2016 and over incompatible formats and higher time computing. Few people want to learn 400 in 2017.18 to compute. These have an indirect a language that can talk only to legacy Capital One’s move from waterfall bearing on run-the-bank costs. technologies, and coders familiar to Agile software development The data collected by banks must with COBOL can be well into their 50s helped cut the time to build new 15 also be analyzed and deployed to or 60s, presenting significant skills application infrastructure by 99%. facilitate better decisions.12 However, shortage challenges in maintaining DevOps’ automation and continuous 16 due to legacy technology, banks are older technologies. Although integration of new code helped speed unable to leverage access to this data upgrades are available, they are still the bank’s development cycles, and for analytics and insights and improve not fit enough to compete or converge releases occurred with increased customer experiences. with systems of this digital era. frequency and higher reliability.19 3. Manual data migration 5. Security concerns 7. Cultural shift Currently, when a bank decides to Customer data in any form is sensitive Banks need to undergo a perception move data to a new system — whether for banks. The major cloud vendors, change. They must be ready to build it’s in the cloud or on-premises — a such as Google and Microsoft, have a culture that is cost-conscious, new project team is set up and works outstanding security expertise, and customer-conscious and in a silo. Data is migrated from the data all are certified compliant with federal efficiency-conscious.20 This is easier lake into the new system, or a new data governance standards. However, said than done, as many systems, connection is created between them. for years banks have delayed and tried processes and people have grown Because of the silos in the bank, other to avoid the issue of modernizing their with the bank. A Boston Consulting teams using the data may be unaware infrastructure. While they agree that Group assessment revealed that of this migration or new connection. cloud infrastructure offers them the among companies that underwent a ability to better cater to the needs of digital transformation, the number of In many instances, multiple teams the digital customer, they are reluctant want the same data and follow a profitable enterprises was five times to adopt it until they are convinced of higher among those that focused on a similar process without getting the safety and security of their data. connected to each other. This leads cultural shift compared with those that They do not have a clear strategy that did not.21 to duplication of data, with multiple will help them quickly adopt cloud. External Document © 2021 Infosys Limited The bank generates and moves (AWS) and Microsoft Azure. How can banks benefit terabytes of data each day. Each Second, it focuses on the data process of data ingestion took eight from cloud? management issue. The platform to 12 weeks to deliver and was Moving data and applications to the enables automated ingestion of data cumbersome, time-consuming and cloud can save banks money. Some versus the manual and siloed approach counterproductive. say it could cut IT costs by as much as previously followed by banks. It also 75%.22 A large global bank that Infosys provides a platform and interface to partnered with to solve this problem Ushering in the Infosys ingest data centrally. estimated that it could save 50% open source data The platform can replicate data of costs overall by adopting cloud. from on-premises to a multi-cloud Some areas in the bank expect to save management platform environment, while supporting batch 90% of their costs in the transition. To solve this problem, Infosys and near-real-time movement. It was To achieve this, the bank is working built an open source, petabyte-scale built with the purpose of enabling with Infosys to build a multi-cloud data multi-cloud data ingestion and guaranteed data delivery at scale. management system that interfaces management platform, part of The data management platform with Google, Amazon and Microsoft Infosys Cobalt. It is a metadata- carries out various functions, removing clouds to migrate its data.23 driven data management ecosystem the need to use multiple types of Cloud is pivotal in architecting the that has been designed to meet an software for each function and saving bank of the future. Banks can benefit organization’s current and future data licensing costs. from the move to the cloud not only by delivery requirements. The platform The platform acquires, ingests reducing costs, but also by capitalizing allows businesses or functions within and transforms data in the on cloud’s computing capabilities, banks to move data from a source to following stages: its ability to scale IT solutions and a destination in a defined format at an its reliability. In fact, moving to the agreed frequency. 1. Acquire cloud can make banks more like the The platform was built to solve a host fintech upstarts with which they Banking transactional data is of banking issues, beginning with data increasingly compete.