SURVIVABLE CLOUD DATA STORAGE AND TRANSPORT

• US 20150248431A1
• Filed: 05/14/2015
• Published: 09/03/2015
• Est. Priority Date: 08/02/2010
• Status: Active Grant

• Current Assignee: Spatial Digital Systems Inc
• Original Assignee: Spatial Digital Systems Inc
• Inventors: Chang, Donald C.D., Chen, Steve K., Lee, Juo-Yu
• Granted Patent: US 10,120,873 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/1727 Details of free space management performed by the file system saving storage space on storage systems G06F3/0608; management of blocks in storage devices G06F3/064
• CPC Class Codes: G06F 16/1744 using compression, e.g. sparse files
• CPC Class Codes: G06F 21/6209 to a single file or object, e.g. in a secure envelope, encrypted and accessed using a key, or with access control rules appended to the object itself

Abstract

Presented is a multi-channel data process to utilize wavefront multiplexing for data storage and data stream transport with redundancy on cloud or in a distribution network. This processing features additional applications for multi-media recording and data communications via transponding platforms including satellites, unmanned air vehicles (UAVs), or others for better survivability and faster accessing. Multiple concurrent data streams are pre-processed by a wavefront multiplexer into multiple sub-channels or wavefront components, where signals from respective data streams are replicated into sub-channels. These replicated data streams are linked via a unique complex weighting vector (amplitude and phase or their equivalents), or “wave-front”, which are also linked by various spatially independent wavefronts. Additionally, probing data streams are embedded and linked via some of the independent wavefronts. Aggregated data streams in sub-channels are unique linear combinations of all input data streams. Generating redundancy for stored or transported data are options in the WF muxing process. Post-processing via a wavefront de-multiplexer converts various sub-channel data streams back to the original data streams, utilizing the embedded known priori data streams as additional knowledge for reducing degrees of freedom to restore orthogonality or independency among multiple wavefronts.

Claims

1. A writing method of optimizing data for distributed storages;

the writing method comprising acts of

transforming multiple (M) input data sets concurrently into multiple (M+N) wavefront components (WFCs) as output data sets via a wavefront muxing processor, where both N and M are positive integers;

identifying multiple selected data storages wherein the data storages distributed in user'"'"'s storage devices and cloud space; and

storing the WFCs to said selected storage site.

2. The writing method for distributed storage of claim 1, wherein the data writing system further comprising segmenting one data stream into said multiple input data sets.

3. The writing method for distributed storage of claim 1, wherein said wavefront muxing processor outputting a first weighted sum of all the inputs as a first output, and outputting a second weighted sum of all the inputs as a second output,

wherein a first input of said wavefront muxing processer contributing to a first aggregated numerical value of the first weighted sum, and concurrently contributing to a second aggregated numerical value of the second weighted sum.

4. The writing method for distributed data storage of claim 1, wherein an act of generating redundancy in stored data further comprising using less number of said input ports to said wavefront muxing processor than the number of said output ports.

5. The method generating redundancy in stored data of claim 4 further comprising an act of connecting at least one of said inputs of the wavefront muxing processor to a known a prioi data stream.

6. The method generating redundancy in stored data of claim 4 further comprising a method of connecting at least one of said inputs of the wavefront muxing processor to ground.

7. The method of distributed data storage of claim 1, wherein the wavefront muxing processor performing a mathematical function of an orthogonal matrix transform.

8. The method of distributed data storage of claim 1, wherein the wavefront muxing processor performing a Hadamard transform.

9. The method of distributed data storage of claim 1, wherein the wavefront muxing processor performing a Fourier transform.

10. The method of distributed data storage of claim 1, wherein the wavefront muxing processor performing a mathematical function of a non-orthogonal and a full rank matrix transform.

11. A reading method of retrieving data from distributed storages on cloud comprising acts of

identifying multiple (M+N) selected data storages in user'"'"'s storage devices and cloud space;

retrieving no less than M wavefront components (WFCs) from said M+N selected storage sites; and

transforming the retrieved M wavefront components (WFCs) as input data sets to M output data sets concurrently via a wavefront demuxing processor;

where M and N are positive integers.

12. The reading method for distributed storage of claim 11, wherein the data reading system further comprising de-segmenting said multiple output data streams into one output data set.

13. The reading method from distributed storage of claim 11, wherein said wavefront demuxing processor outputting a first weighted sum of all the inputs as a first output, and outputting a second weighted sum of all the inputs as a second output,

wherein a first input of said wavefront demuxing processer contributing to a first aggregated numerical value of the first weighted sum, and concurrently contributing to a second aggregated numerical value of the second weighted sum.

14. The reading method from distributed data storage of claim 11, wherein an act of retrieving from stored redundant said M+N data sets further comprising using no less than M said input ports to said wavefront demuxing processor connected to said M output ports.

15. The reading method from stored redundant data of claim 14 further comprising a method of connecting at least one of said outputs of the wavefront demuxing processor to a known a prioi data stream.

16. The reading method from stored redundant data of claim 14 further comprising a method of connecting at least one of said outputs of the wavefront demuxing processor to ground.

17. The reading method from distributed data storage of claim 11, wherein the wavefront demuxing processor performing a mathematical function of an orthogonal matrix transform.

18. The reading method from distributed data storage of claim 11, wherein the wavefront demuxing processor performing a Hadamard transform.

19. The reading method from distributed data storage of claim 11, wherein the wavefront demuxing processor performing a Fourier transform.

20. The reading method from distributed data storage of claim 11, wherein the wavefront demuxing processor performing a mathematical function of a non-orthogonal and a full rank matrix transform.