The Assessment: Our client is a large federal agency with siloed data. Throughout the agency, people had to rely on manual processes that were difficult to maintain due to lack of data centralization. Parts of the organization that needed regular access to each other’s data had to constantly play catch-up to keep as current as possible.
The DSS Solution: DSS began to address these pressing issues by using vendor-supplied and custom tools that monitor and collect data from legacy, current, and new storage infrastructure compute and media devices supporting many teams throughout the agency. Doing so led us to create DIAMOND our proprietary configuration management database (CMDB) illustrated in Figure 1 below. DIAMOND collects data from 37 application APIs and numerous Oracle and Postgres third-party databases to tabulate real-time data on 12,000 servers, 3,000 file shares, 30,000 storage volumes, 95 clustered storage containers, 30 switches, and 45 arrays.
The DSS Difference: Unlike the old days of slow-to-update, siloed data, DIAMOND now provides the client with a single pane of glass that provides a snapshot of all systems in the IT infrastructure. Through DIAMOND, we provide continuous monitoring and analysis of key performance indicators (KPIs), including, but not limited to storage consumption use and trends, reclamation of underused storage capacity, identification of cost savings, capacity management to support the load growth, storage expansions, and re-balancing of resources, performance hot spot identification for risk mitigation consideration, and performance metrics reviews for compliance with SLAs. Through DIAMOND’s Chargeback & Showback features, we integrate business architecture and IT architecture for internal billing. Figure 1 below shows current data sources that have automation built in, polling the infrastructure for accurate and real-time information.
Figure 1: DIAMOND Centralizes and Automates Data Lookups from 37 Sources
DIAMOND is a modular data aggregation and analytics application composed of several functional layers. These functional layers work together to provide the USPTO with an exceptional technology business management (TBM) application. The approach we used in creating DIAMOND can be applied to a variety of data analytics systems, including CMDB, asset management, and physical environment data. Figure 2 depicts DIAMOND’s systems architecture.
DIAMOND: A Modular Data Aggregation and Analytics Application
DIAMOND also enables cost reductions by providing utilization metrics across operating systems and architectures, which helps stakeholders understand the current total cost of ownership (TCO) better as well as future storage needs. Because of this, we were able to assist with an agency-wide storage reclamation effort wherein we examine file systems to see where storage can be optimized to 85% capacity. When we encounter a server that has been given an excess of storage, i.e., its storage capacity is under 85%, we re-provision it. By doing so, we are able help the client achieve savings in that they pay less for storage. To date, the storage reclamation effort has realized a cumulative cost avoidance for storage of $186K monthly/$2.24M annually from the storage invoice since the effort began in 2019. Please see Figure 3 below, which depicts storage savings for FY 22 and FY 23.
Figure 3: Cost Savings Due to Storage Reclamation, FY 22-FY 23
DSS software and systems engineers develop, maintain, and provide ongoing technical support for the frontend and backend of DIAMOND. We follow an Agile development process based in Scrum in a DevSecOps environment that allows us to maintain DIAMOND while simultaneously developing and integrating new features.
Due in large part to the success they have enjoyed because of DIAMOND, the client engaged DSS to build an enterprise-wide database using ServiceNow in November of 2023. By re-using and extending the existing code, APIs, and infrastructure already in place with DIAMOND, the delivery milestones along the path to completion have been significantly accelerated.
