Browse Prior Art Database

Solution Search Technique for Generating Recommendations for Heterogeneous Combinations Servers and Features

IP.com Disclosure Number: IPCOM000236736D
Publication Date: 2014-May-13
Document File: 4 page(s) / 57K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for representing and searching for valid computer hardware configurations that satisfy a set of requirements. The search design is independent of specific hardware architectures.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 48% of the total text.

Page 01 of 4

Solution Search Technique for Generating Recommendations for Heterogeneous Combinations Servers and Features

A method is needed to find server hardware recommendations (solutions) that are sufficient to run a given application workload. The recommendable hardware types span different brands/architectures of servers. Programs/scripts called Sizing Guides exist to calculate predicted system requirements based on answers to a questionnaire about a given software application.

From customer or business partner perspective, the current recommendation process is:

1. The customer answers a questionnaire about the intended use of a software application

2. A Sizing Guide specific to that application calculates system requirements 3. A Sizing Engine searches/calculates the list of possible server configurations that satisfy the requirements 4. The Sizing Engine attempts to find the optimal server configuration and recommend it to the customer
5. The customer can order the recommended server hardware

From an internal perspective, the current search solution uses a combination of

1. Documenting server system metrics in several large spreadsheets, with different spreadsheet formats for each brand of server

2. Reading the spreadsheet into memory in a Sizing Engine program 3. Evaluating every server configuration to find recommendations that can support the software workload. Specific hardware architecture and configuration information is programmed into the sizing algorithm.

With the current approach, the format of the spreadsheet must be different for every brand of hardware. It is expensive to update algorithms for new hardware designs. A new version of the sizing software is required for every new set of recommendable servers, so it is expensive to incorporate new recommendable hardware technology. Scalability is an issue because the current design individually processes each combination of hardware to determine whether it can successfully run the given workload.

The novel contribution provides a way of representing and searching for valid computer hardware configurations that satisfy a set of requirements. The hardware and capabilities are identified in a schema-less tree. The tree topology describes hardware configurations (i.e. which features can be used with which servers); therefore, this method allows the search design to be independent of specific hardware architectures. The design finds configurations that satisfy a set of requirements by finding subtrees that contain a set of nodes with capabilities corresponding to the requirements.

With this method, all hardware can be described in one datastore. Sizing Engine software does not have to change every time the server hardware changes. By

1


Page 02 of 4

inserting new systems into an existing tree, new recommendable hardware can be added "on the fly", without the need for new software changes.

The basic idea of the solution is to search for combinations of products (i.e. servers) and fe...