Hana Multi-Temperature Data Strategy

Each layer of the Hana architecture is based on the locality of the data, with capacity and performance requirements usually handled in parallel. The first layer in the data temperature model is for "hot" data that is accessed frequently and has the highest performance requirements.

The second layer is for "warm" data, i.e. data that is rarely accessed and has lower performance requirements, but which must remain at the core of the database so that operations can continue.

The third and final level is for "cold" data, i.e. information that is accessed sporadically with low performance requirements.

Implementing a multi-temperature data strategy for Hana can help reduce total cost of ownership (TCO) while maintaining an organization's long-term data growth strategy. Hot data for Hana resides in memory.

As data growth takes place in this layer, more physical storage is required, which in turn consumes more power and cooling, resulting in higher costs for operating the database.

In the event that some of the data in the hot tier is not frequently used, this provides an opportunity for cost savings by moving it to the warm tier. The warm tier consists of the following data temperature management techniques.

Paged Attributes

This parameter only applies to S/4 and SoH. This allows the columns of a table (only for column tables) to be loaded and unloaded between memory and persistent storage as required. This is implemented by the Data Aging Framework provided by NetWeaver Abap.
Dynamic tiering

A native Hana technology that makes it possible to place tables in the extended storage and thus increase the capacity of the warm data volume.
Extension Nodes

Based on the Hana scale-out architecture, a node in a Hana distributed system can be specified for the management of warm data. Data that needs to be moved from a hot tier to the warm tier in this scenario is redistributed to the extension node.

An extension node makes it possible to store a larger amount of data (100 percent of the dram size) than the normal Hana size requirements (50 percent of the dram size).

This hot data method can be implemented natively or with BW/4 Hana and SAP Business Warehouse on Hana.
Native Storage Extension (NSE)

A native and built-in warm data store that allows data that is accessed less frequently to be managed by accessing it from the hard disk instead of in memory.

With poor data management, the cost of developing, implementing and managing SAP can quickly get out of hand. Given the increasing digitalization of companies, every minute wasted can potentially be very expensive.

Pure Storage is an application development partner and participant in the SAP Partner Edge Program, offering solutions for Hana applications. With high performance and resilience, Pure can significantly reduce operating costs.

With Pure's solutions, SAP can be accelerated up to tenfold on the basis of a traditional relational database. If the existing customer decides to switch to Hana, the same certified arrays can be used. No further purchases are necessary.

With a growing number of Hana users, Pure shows how simplicity and performance can save millions of dollars, not only in the data center, but also at the application level. The right storage platform can make all the difference to Hana performance.

Pure began by delivering all-flash performance with best-in-class data reduction for SAP AnyDB and Hana. This enables fast, smooth business processing and efficient all-flash storage to deliver performance and cost savings at the same time.

Pure demonstrated the use of Hana's Native Storage Extension (NSE) for managing "warm" data, coupling the new Direct Memory Cache based on Intel Optane.

The result is 90 percent of the performance of in-memory Hana at a fraction of the cost, dramatically improving TCO without a noticeable increase in latency.

Building on the NSE and direct memory cache innovation, Pure-Flash-Array-X offers reduced response times for analytical queries and reporting on data using warm data management.

Flash Array X increases analytical query performance for Hana by up to 50 percent over the previous generation when using warm data management.

Of course, nothing is as fast as 100% execution in memory. But this performance comes at a high cost for both Hana licenses and compute infrastructure - which can be millions of dollars for a company. In comparison, Flash Array X as part of the Evergreen Gold subscription boosts performance for these types of workloads at no additional cost.

As with previous offerings for existing SAP customers, Pure worked closely with SAP to ensure that Flash Array-X is fully certified as an enterprise storage solution.