Earth Science Data Systems Level of Service Model

Overview

Data archived at NASA’s Earth Science Data Systems (ESDS) data repositories are provided long-term, high-quality data stewardship that includes more services than just data archival and distribution.

However, each individual data product cannot and likely should not have all the same services applied. The typical goal is to apply the appropriate amount of services to a data product to optimize data usability and access while also balancing the resources, time, and effort required from both the data provider and the data repository.

The Level of Service model is designed to bring consistency to service determinations for data across all NASA ESDS data repositories and to improve communication of the data archival roles and responsibilities, making it easier for data producers and data users to know what to expect with respect to data stewardship of a given data product.

The term "Level of Service" refers to the quantity and types of services applied to data during archival and preservation. Services include the activities related to data ingest, storage, documentation, distribution, assurance of usability and discovery, user support, outreach, and preservation.

The process of determining what services to provide a data product has previously been carried out independently at each of NASA's Distributed Active Archive Centers (DAACs). Therefore, data producers interacting with different data repositories may not have received the same services for similar data products. This has led to data producer confusion and frustration and highlighted the need for creating a common model. The primary goals of the Level of Service Model are to:

• Provide a framework for clearly communicating information about data stewardship at NASA ESDS data repositories.
• Use consistent terminology so that data producers interacting with multiple data repositories can understand the process no matter which data repository they interact with.
• Present clear definitions and service level descriptions so that all data producers and users can understand how services are assigned and have realistic expectations of data repository effort.
• List ESDS common data stewardship requirements by service level. Data repository-specific services are not included in the model and would need to be specified separately. 
• Improve enterprise-wide transparency, efficiency, and cost-effectiveness of the data stewardship process.

How is Service Level Assigned?

Service levels are assigned based on an assessment of the best investment of NASA resources to maximize the overall science value of NASA-funded data. The data processing level and data product category assigned by NASA provide a framework for determining the appropriate level of service. However, science needs, data age, and level of effort may drive increases or decreases in assigned service level. Level of effort, and thus the practicality of specific services, can be greatly affected by the comprehensiveness and clarity of the supplied documentation and metadata, as well as the degree to which the data follow current NASA data standards. Documentation, metadata, and standards compliance may also allow a data center to provide selected services from higher levels at little to no additional cost.

NASA's data processing level identifies the nature of the data, from raw instrument data (Level 0) to gridded variable data products (Level 3) to model output (Level 4).

There are six NASA data product categories:

• Science Mission Products
• NASA-funded Data Products
• Facility Created Products
• NASA Research Community Products
• NASA Applied Science Products
• Non-NASA Research Community Products

More information about the data product categories can be found in the Level of Service model document (PDF).

The Level of Service Determination table below identifies which of three service levels (basic, standard, or comprehensive) generally apply to data of a given processing level and data category. The second table shows service level descriptions. Although a near real-time (NRT) data product may have a comprehensive service level designation due to its product category and processing level, it may not obtain all comprehensive level services as some may not apply and others may not be realistic.

Table 1. Typical Level of Service Determinations

Service Level Guide for Data Producers

The intent of the ESDS Level of Service model is to describe the services to be provided for a given data product, as opposed to suggesting how those services will be implemented. This leaves room for data repository personnel to determine the best way for them to meet the assigned service level. Each data repository has unique capabilities and services that are designed to meet the needs for their particular data producer and user communities. These capabilities may allow the repository to offer services not specifically listed here and/or offer selected additional services beyond those that are required.

The Level of Service model covers services in six categories, including Ingest, Check, and Archive; Distribution and Access; Documentation; Discoverability and Usability; User Support and Outreach; and Preservation. The requirements directly relevant to data producers (in the categories of Data, Documentation, and User Support and Outreach) are summarized in Table 2. These lists demonstrate what you can expect a DAAC to do with respect to your data product as far as handling data, documenting data, and supporting data into the future.

The basic level requirements are designed to provide a minimal level of data stewardship that makes the data findable and accessible. Even at the basic level there is a high bar on data requirements that ensures appropriate data stewardship. The additional requirements for standard and comprehensive services build on these baseline requirements, increasing the quality of data care and support. Particularly for the comprehensive service level, not all services are appropriate (or even possible) for any given data product, but data producers can expect that the data repository will provide as many of these as are appropriate and reasonably feasible. 

These requirements apply only to upcoming/future datasets at this time. Requirements for already published and supported data have not yet been determined.

Table 2. Service Level Descriptions