Utility-based archiving

• US 20040243774A1
• Filed: 06/16/2004
• Published: 12/02/2004
• Est. Priority Date: 06/28/2001
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Horvitz, Eric J.
• Application Number:
  US10/869,683
• Publication Number:
  US 20040243774A1
• Time in Patent Office: Days
• US Class Current: 711/161
• CPC Class Codes: G06F 16/10 File systems; File servers
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99955 Archiving or backup

Abstract

A system and associated methodology is provided that is adapted to infer what to do with an item, and more particularly whether to archive and/or keep active an item in a more active, easy-to-access store based upon a cost-benefit analysis. The cost-benefit analysis determines the overhead associated with keeping the item active (e.g., not archiving it) versus the gains in connection with having quick and easy access to the item. The cost of maintaining an item in an active state is measured in terms of the size of the item which, in turn, affects the amount of space needed to store it. The benefit of keeping the item active is measured in terms of a probabilistic determination describing how a user will access the item in the future, which is a reflection of the utility of the item in an accessible state. The invention leverages notions of temporal sensitivity of the likelihood that an item will be needed in the future such that determined values and inferences can be dynamically updated over time. Items having a small probability of being accessed again after an initial review are categorized as one-shot items.

Claims

1. A system that stores an item comprising:

a component that utilizes a formal probabilistic analysis to infer whether to store the item in an active item store or an archive item store.

2. The system of claim 1, the component determines a value density for the item based upon a probability of user access of the item.

3. The system of claim 2, the value density is the probability of access of the item given evidence divided by the size of the item.

4. The system of claim 1, the component infers to store the item in the archive item store if a probability of user access is less than a threshold value.

5. The system of claim 2, the component infers to store the item in the archive item store if the value density is less than a threshold value.

6. The system of claim 2, the probability of user access of the item is time dependent.

7. The system of claim 6, the time dependent probability of access decays over time.

8. The system of claim 7, the decay is based on at least one of policies, and information regarding the item, a user, and extrinsic data.

9. The system of claim 6, the time dependent probability of access increases over time based on user re-access of the item.

10. The system of claim 1, further comprising a user interface that facilitates user control.

11. The system of claim 10, the user interface facilitates control over policies that affect probabilistic determinations.

12. The system of claim 11, the user provides an amount to reduce the probability of access by after the item is accessed via the user interface.

13. The system of claim 11, the user interface facilitates reducing the probability over time by a selected amount after an amount of time selected by the user.

14. The system of claim 11, the user interface enables a user to control a threshold probability of access more than once that defines a one-shot item.

15. The system of claim 11, the user controls the amount by which to increase the probability when the item has an attachment.

16. The system of claim 10, the user interface facilitates control over an archiving rule.

17. The system of claim 11, the user controls a threshold value density and an item with a value density less than the threshold value density is stored in the archive item store.

18. The system of claim 11, the user interface provides for the user to select a minimum age of the item that is stored in the archive item store.

19. The system of claim 11, the user interface facilitates user selection of whether to archive one-shot items after they are read.

20. The system of claim 10, the user interface provides user control regarding item discard protocols.

21. The system of claim 20, the item discard protocol is a protocol that discards an item from a user selected sender.

22. The system of claim 10, the user interface enables user control over memory utilization protocols.

23. The system of claim 22, the memory utilization protocol is a user controlled number of items stored in the active item store.

24. The system of claim 22, the memory utilization protocol is a user selected percentage ceiling of active space as the active item store.

25. The system of claim 1, further comprising a learning system that analyzes at least one of the items, user data, and extrinsic data to adapt inferences performed via the formal probabilistic analysis.

26. The system of claim 25, the learning system utilizes classifiers that consider distinctions between at least one of an address, an attachment, header information, and a body structure of the item.

27. The system of claim 1, further comprising an optimization component that maximizes the total value density, the optimization component selects items with the greatest value densities to store in the active item store and examines remaining items to determine which fit into a residual amount of space within the active item store.

28. A method for determining how to store an item comprising:

utilizing formal probabilistic analysis to determine probability of user access of an item;

determining a value density of an item based on the probability of user access; and

inferring whether to store the item in an active item store or an archive item store based on at least one of the probability of user access and the value density of the item.

29. The method of claim 28, further comprising storing the item in the archive item store based on at least one of the probability and the value density being less than a threshold value.

30. The method of claim 28, further comprising decreasing at least one of the probability of user access and the value density over time based on at least one of policies, and information regarding the item, a user, and extrinsic data.

31. The method of claim 28, further comprising increasing at least one of the probability of user access and the value density over time based on a user re-accessing the item.

32. The method of claim 28, further comprising controlling policies that affect the formal probabilistic analysis via a user interface.

33. The method of claim 32, further comprising selecting at least one of a probability reduction value for after an item is selected, a time based probability reducing amount, a threshold value to identify a one-shot item, and a probability increasing value corresponding to an attachment for the item.

34. The method of claim 28, further comprising controlling an archiving rule by selecting at least one of a threshold density value below which items are stored in the archive item store, a minimum duration of time prior to storing an item in the archive item store, and whether to archive one-shot items subsequent to reading.

35. The method of claim 28, further comprising controlling an item discard protocol via a user specifying senders of items to discard.

36. The method of claim 28, further comprising controlling a memory utilization protocol via a user selecting a number of items to store in the active item store and a percentage ceiling of active space as the active item store.

37. The method of claim 28, further comprising adapting the performed inferences via machine learning.

38. The method of claim 37, learning utilizes classifiers that distinguish between items, the classifiers are built by watching a user re-access email over time and automatically training a filter that assigns an incoming message at least one of a time until future access and a probability of access.

39. The method of claim 28, further comprising archiving an item to minimize latencies via maximizing the overall value densities of the items stored in the active item store.

40. A system that archives items based on utility comprising:

means for performing formal probabilistic analysis upon an item;

means for determining value density of the item based on the results from the formal probabilistic analysis; and

means for storing the item in an active item store or an archive item store based upon at least one of the formal probabilistic analysis and the value density.