Служба иммиграционного контроля США ищет подрядчиков для модернизации централизованного хранилища данных

ICE procurement reveals data lake underlying immigration enforcement analytics U.S. Immigration and Customs Enforcement (ICE) is seeking contractor support for the Enforcement and Removal Operations (ERO) Data Lake, a centralized analytical environment that integrates information from multiple immigration enforcement systems for reporting, governance and analysis.\ A Request for Information (RFI) issued by ICE’s Office of the Chief Information Officer shows the platform operating alongside ICE’s long-standing reporting environment using AWS GovCloud, Databricks, Databricks Unity Catalog and Qlik Sense. The procurement offers a rare glimpse beneath operational enforcement applications, revealing how ICE standardizes and analyzes information from multiple systems without merging the operational databases themselves. The requirements show ICE maintaining and modernizing a data integration architecture that predates its newer Palantir-based ImmigrationOS and ELITE systems. The architecture allows information from separate operational databases to be analyzed together without merging the source systems. The contractor would support operations, maintenance and modernization across ERO systems, including the ERO Data Lake, ICE Data Reporting Systems (IDRS), data engineering, analytics and reporting. Among the requirements is designing and maintaining extract, transform and load (ETL) pipelines that standardize information from operational systems before moving it into the data lake for analysis. ICE also identifies Databricks and its Unity Catalog technology for data integration, transformation, quality and governance. Qlik Sense provides business intelligence and visualization capabilities, while AWS GovCloud provides cloud infrastructure for government workloads. From data warehouse to cloud data lake ERO has maintained a centralized enforcement data environment for years through IDRS, also known in government records as ICE Integrated Decision Support (IIDS). ICE described the architecture during a 2021 industry engagement as having four major components. Operational systems supply source data. ETL processes extract and normalize the information. A centralized repository stores current and historical data. Analytical applications then allow authorized users to query and report on it. A Government Accountability Office (GAO) report describes IIDS as an ERO data warehouse populated with ENFORCE information related to the investigation, arrest, booking, detention and removal of people encountered during immigration and criminal law-enforcement operations conducted by ICE and Customs and Border Protection. GAO itself analyzed individual-level IIDS records containing fields including A-numbers, names, dates of birth, person identifiers, citizenship and criminality information. Department of Homeland Security (DHS) privacy records provide more detail about what analysts could do with the information. They describe ERO personnel using IIDS for predefined and ad hoc analysis of enforcement data involving encounters, arrests, bonds, detainers, detention and criminal and removal cases. DHS records say authorized users could also query individual records using personal identifiers. The RFI indicates that this longstanding integration function is being maintained and modernized around a cloud-based data-lake architecture. Connecting separate enforcement systems A data lake differs from a conventional database. Rather than being designed primarily around a particular operational application, it can receive large quantities of information from different sources and make that information available for transformation, analysis and other downstream uses. ICE’s requirements accordingly call for contractors to maintain extract, transform and load pipelines into the ERO Data Lake, perform data-quality work, administer Databricks and Unity Catalog, support analytics and reporting and help migrate or retire legacy systems. During the Secure Communities program – which checked the fingerprints of people arrested by local police against immigration databases to help identify removable noncitizens – ICE used IIDS to connect information originating in systems that were not themselves integrated. One was the Alien Criminal Response Information Management (ACRIMe) system used by ICE’s Law Enforcement Support Center. Another was ENFORCE, ICE’s suite of operational immigration enforcement applications. Because ACRIMe was not integrated with ENFORCE, ICE used IIDS to link Law Enforcement Support Center identification records from ACRIMe with corresponding ENFORCE information already residing in the IIDS warehouse, automating what ICE described as the process of pairing identification data with enforcement information about a subject The example illustrates a fundamental feature of the current architecture. Government databases do not necessarily have to be combined into one enormous operational system for their information to become analytically connected. Data can instead be extracted from separate systems and brought into a common environment where it can be standardized and analyzed together. ICE’s older IIDS architecture shows that such an environment can also be used to match records originating in otherwise separate operational systems. The RFI suggests ERO is continuing that model with considerably newer technology. ICE also wants support for data governance, data quality, user enablement, DevSecOps, release engineering and modernization of older systems. The procurement suggests ICE is preserving the underlying architecture developed through IIDS while replacing legacy components with cloud-native technologies. Rather than consolidating operational systems into a single platform, the agency continues to rely on a centralized analytical layer that brings together enforcement data for governance, reporting and operational analysis. Article Topics ICE - U.S. Immigration and Customs Enforcement | immigration | procurement | RFI | U.S. Government