Browse Prior Art Database

Disaggregate compute - CH8-2015-0101 Re-submission

IP.com Disclosure Number: IPCOM000245592D
Publication Date: 2016-Mar-21
Document File: 2 page(s) / 28K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method and system to automatically distribute blocks of software code (code snippets) of an application onto various compute nodes based on identifying such code blocks during load- or runtime. This allows conversion of applications, libraries, etc. automatically into distributed components available as software as a service (SaaS).

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 2

Disaggregate compute -

Disclosed is a method and system to automatically distribute blocks of software code (code snippets) of an application onto various compute nodes based on identifying such code blocks during load- or runtime. This allows conversion of applications, libraries, etc. automatically into distributed components available as software as a service (SaaS).

Background Code snippets are blocks of reusable code that can be inserted during programming of an application. Snippets can be very simple, such as conditional statements or loops, but also complex blocks, such as methods or classes. Various application development environments [1]
[2] incorporate a snippet manager that facilitates the usage of code snippets in order to speed up application development. While code snippets are a common and important technique in application development, and even insertion of code snippets at run time has been disclosed [3] they loose their relevance as soon as they are integrated into an application. So during runtime for example code snippets are not identified or treated as special entity. But as code snippets represent common tasks within an application, task distribution based on snippets to different processing nodes or special processing nodes can improve performance. For special snippets like data base queries the distribution onto special execution engines was already disclosed [4] and is used e.g. in massive parallel processing systems.

Description The solution is a system comprising a snippet identification unit, a snippet manager, and various processing nodes where code snippets are automatically identified during load or runtime through code profiling, or source code annotation, or the beginning of a library call, or run-time profiling, etc. The snippet identification unit identifies suitable code snippet...