MCA Data Science — free MCQ practice

1. The definition is short: "Horizontal scaling is increasing the capacity of production output through the addition of contributing resources." 2. The word that matters is addition. Nothing is improved; more of it is added. 3. The framework calls this "data-intensive parallel processing" and says it "is at the heart of the Big Data revolution", achieved through "the maturation of techniques for distributed processing across independent resources, or nodes in a cluster". 4. Improving existing components is the definition of vertical scaling, the other method. 5. Sampling before analysis is a statistical practice the framework discusses under data science, not a scaling method. Reapplying an entity for a different purpose is its definition of reusability. 6. The framework states the ideal that horizontal scaling aims at: "twice the data can be processed in the same amount of time using twice the number of data nodes", which it calls linear scalability. _Source: NIST SP 1500-1r2 "NIST Big Data Interoperability Framework: Volume 1, Definitions" (Oct 2019), p.10 section 3.1 "Big Data Definitions"_

1. The definition is one line: "Data locality refers to the data processing occurring at the location of the data storage." 2. The idea is to move the computation to the data rather than the data to the computation, which is what makes distributed processing across a cluster efficient. 3. Note that locality here is about the relationship between processing and storage, not about geography. Nothing in the definition mentions regions or jurisdictions. 4. The delay in processing or availability is the framework's definition of latency, a neighbouring term in the same list. 5. The tendency of data structures to change over time is its definition of volatility, another term from the same list. _Source: NIST SP 1500-1r2 "NIST Big Data Interoperability Framework: Volume 1, Definitions" (Oct 2019), p.6 section 2 "Terms and Definitions"_

1. The definition reads: "Schema on read is the application of a data schema through preparation steps such as transformations, cleaning, and integration at the time the data is read from the database." 2. So the schema is not absent, it is deferred. The structure gets applied at read time rather than at write time. 3. The three preparation steps the definition names are transformations, cleaning and integration. 4. Defining the schema before anything is written is the opposite arrangement, the one used by the relational model the framework describes as the older default for business systems. 5. This connects directly to the framework's point about unstructured data. When records do not share a consistent structure, forcing a schema at write time is expensive so the structure is imposed later, by whoever reads the data for a particular purpose. _Source: NIST SP 1500-1r2 "NIST Big Data Interoperability Framework: Volume 1, Definitions" (Oct 2019), p.7 section 2 "Terms and Definitions"_