Richard Myron Swift

US Patent:

20020120724, Aug 29, 2002

Filed:

Feb 23, 2001

Appl. No.:

09/792805

Inventors:

Christian Kaiser - Cupertino CA, US
Richard Swift - Sunnyvale CA, US

International Classification:

G06F015/177

US Classification:

709/221000, 709/220000

Abstract:

A mechanism for reconfiguring a server without incurring any server down time is disclosed. At startup time, a server constructs a set of configuration data structures based upon a set of configuration information, and stores the data structures into a portion of the server's memory space. The server also updates a current configuration pointer to point to the configuration data structures to establish the data structures as the server's current configuration. Then, during normal operation, the server receives a reconfiguration signal. In response to this signal, the server constructs a new set of configuration data structures based upon a modified set of configuration information, and stores the new data structures into another portion of the server's memory space. Then, the server updates the current configuration pointer to point to the new configuration data structures to establish the new configuration data structures as the current server configuration. By doing so, the server changes its configuration. This change is implemented without shutting down and restarting the server.

US Patent:

20170139914, May 18, 2017

Filed:

Nov 18, 2015

Appl. No.:

14/945318

Inventors:

- Redwood Shores CA, US
Han Shum - Los Gatos CA, US
Richard Swift - Sunnyvale CA, US
Bharathi Sivaramakrishnan - San Jose CA, US

International Classification:

G06F 17/30
H04L 12/58

Abstract:

Techniques are herein described for creating a scalable IMAP4 compliant email system using a NoSQL database and a distributed full text search engine. Data for each email message is stored in multiple tables to avoid storing redundant data unnecessarily. However, a full text search index is created based on a single table as if the index refers to a single table. In embodiments herein described, the single index is created on the fields of a message metadata table with virtual fields added to the table that are derived from the message content. During this process, data is pulled from a message table in “blob” format and broken down into corresponding fields and data items, so the data items may be converted and placed in the proper virtual fields for index creation. Each blob section that is converted is cached, so the same blob section does not need to be converted multiple times. After index creation, the index may be used to search for emails based on metadata and data within the body of the email.

US Patent:

20170068732, Mar 9, 2017

Filed:

Sep 3, 2015

Appl. No.:

14/845121

Inventors:

- Redwood Shores CA, US
Richard Swift - Sunnyvale CA, US

International Classification:

G06F 17/30

Abstract:

A method includes storing a copy of information, corresponding to a data set, in each of a plurality of data formats, and determining one or more portions of a search criterion to be associated with one or more sets of one or more characteristics. For each portion, based on the associated characteristics, a search is executed on a copy of information stored in a particular data format using a corresponding search algorithm to obtain a set of intermediary search results. A set of final search results corresponding to the search criterion is determined based on two or more sets of intermediary search results.