InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Smart Data Driven Dynamic Caching Algorithm

IP.com Disclosure Number: IPCOM000225805D
Publication Date: 2013-Mar-07
Document File: 6 page(s) / 75K

Publishing Venue

The IP.com Prior Art Database


Disclosed is a method for maintaining e-commerce website content that allows business users to change any entity on a given page at any time. A business user can define dynamic or static rules in each area of the page that controls what to show, when to show it, and perhaps target a specific audience. This invention describes a pattern giving a business user complete control over an application interface without thinking or navigating around cache policies and a stringent non-changeable cache rule.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 26% of the total text.

Page 01 of 6

Smart Data Driven Dynamic Caching Algorithm

Online e-commerce is an example of an activity where the interests of the Chief Marketing Officer and Chief Technology Officer conflict.

Marketing managers demand a user experience that is engaging, dynamic, relevant, and targeted, and to achieve this the managers want site content to be determinable by flexible business rules. Marketing also highly values giving the marketing manager direct control over these rules, so that the manager can responsively schedule and change the rules without IT becoming a bottleneck .

At the same time, IT wants to achieve high throughput and low response time. Caching is an effective tool to achieve this. A current leading caching technology is

controlled by a combination of Application Protocol Interfaces (APIs) and Extensible Markup Language (XML) configuration files, which determine the caching policy for

the site including caching granularity, cache keys, invalidation keys, and timeouts for cache entries. Policies that create large granularity cache entries, with non-selective cache keys to maximize reuse, and with long time-outs and rare invalidation improve performance; however, such policies can deliver stale or incorrect results if incompatible with the business rules that the marketing manager provides.

Current practice is that IT defines compromise caching policies that assume some amount of personalization and time variation in different parts of a page , and then the marketing manager must only use business rules that stay within those boundaries. If the business rules in effect at a given time are less dynamic and less personalized than the compromise permits, then performance is unnecessarily degraded. On the other hand, if the business rules are excessively dynamic or personalized, then users may be delivered cached content that is out of date or that is appropriate for some other user. Figure 1 provides an example.

Figure 1: Consider two cacheable objects. Entities within these objects can either be individual (entity #1) or shared (entity #2).

If object #1 and object #2 are both cached, then all three entities can be in one of the following states:

¡⁄ "Cacheable" and "consumable" by the parent object; meaning there is one cache record including the entity along with the object

¡⁄ "Cacheable" and "non-consumable" by the parent object; meaning there are two cache records, one for the entity and the other for the object

¡⁄ "Non-cacheable" and "non-consumable" by the parent object; meaning there


Page 02 of 6

is one cache record including object and no cache record for the entity (since if cannot be cached)

Assume that IT defines a cachespec such that entity #2 is "Cacheable" and "consumable". This means that entity #2 cannot contain data that has dynamic behavior. Taking an e-Commerce example into consideration, object #1 might be a "woman's" category page and object #2 might be an "on-sale" category. Both are cacheable objects, b...