Storage asaService
The storage capacity growth trend seen from cloud vendors is moving positively toward consumption-based offerings. Given the need to reduce time and costs, and to increase the ease of IT storage administration, support and maintenance will be moving toward storage as a service (STaaS) and artificial intelligence for IT operations (AIOps).
Storage Solutions
Storage infrastructure vendors provide solid-state arrays (SSAs), hybrid arrays, and software-defined storage (SDS) solutions that are deployed to support primary storage workloads. Storage in terms of modern data warehousing has some criteria.
Advances in cloud-native technologies; managed STaaS offerings; advanced AIOps features; consumption-based, SLA-metric-driven sourcing; and vendor-based asset financing and management programs set the stage for the next phase of the enterprise storage industry.
In-memory Storage
In-memory systems are data warehouse management systems (DBMS) where data is stored within computer memory itself instead of on a physical hard drive. This approach creates faster access and gets high I/O performance of system resources. This is critical for real-time applications like transaction processing and other applications that need to read and write data quickly. Three important factors of performance are memory, processing, and I/O. The physical design of most systems tries to speed up processing by reducing memory/storage access.
One of the approaches to doing so is by duplicating the data within or near a system where processing is performed; e.g., one version is placed in a database for operational OLTP (online transaction processing) systems, and another version in data warehouses for OLAP systems. This is because OLTP is I/O heavy, while OLAP is computationally heavy and requires different optimization routines. In-memory database applications can get you the best of both worlds in a single system. However, make sure that sensitive data is encrypted before storage and that there is regular back-up of data.
An in-memory-based approach can be implemented for column-based storage as well as for relational databases; e.g., Oracle’s TimesTen for in-memory relational DB, SAP HANA for a column-based data warehouse, and Starcounter for OLTP solutions.
