IP Library › Granted Patent US 12,656,966
Granted Patent B2
US 12,656,966 · App. 17/836,727 · Granted Jun 16, 2026

Coordinated allocation of external memory

Inventors: Timothy A. Stabrawa (Lombard, IL); Zachary A. Cornelius (Chicago, IL); John Overton (Chicago, IL); Andrew S. Poling (Lombard, IL); Jesse I. Taylor (Glenside, PA)
Assignee: Kove IP, LLC
G06F3/0632G06F3/0604G06F3/067G06F12/1027G06F2212/683
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,656,966
App. No.
17/836,727
Filed
Jun 9, 2022
Granted
Jun 16, 2026
Kind
B2
Art Unit
2137
USPC
711/172
Abstract

Methods, devices, and executable instructions are provided for allocating external memory. A request to allocate a portion of memory may be received at a first device. A memory appliance and the first client device are connected by a network. A region of memory of the memory appliance may be allocated as external memory prior to receipt of the request to allocate the portion of memory at the first client device. A subset of the region of memory for the portion of memory allocated may be selected at the first client device and in coordination with a second client device on the network, wherein the portion of memory allocated is external to the first client device and yet is primary memory to the first client device. At least the subset of the region of memory may be mapped at the first client device to a virtual address space.

Claims (26)

1 . A non-transitory computer-readable storage medium comprising a plurality of instructions executable with a processor, the instructions including:

instructions executable to receive, at a first client device from a component in the first client device, a request to allocate a portion of memory, wherein a memory appliance and the first client device are on a network, and a region of memory of the memory appliance is allocated as external memory prior to receipt of the request to allocate the portion of memory at the first client device, wherein the external memory is external to the first client device and a second client device, but is primary memory to the first client device and the second client device;

instructions executable to select, by a first client logic within the first client device, in coordination with a second client logic within the second client device and in response to the request to allocate the portion of memory, a subset of the region of memory of the memory appliance, the first client logic being otherwise distinct from the second client logic aside from coordinating with the second client logic in using the memory of the memory appliance; and

instructions executable to map, at the first client device, at least the subset of the region of memory to a virtual address space, wherein data in the portion of the external memory is accessible from the first client device via client-side memory access in which a communication interface of the memory appliance is configured to access the region of memory on the memory appliance,

wherein selection of the subset of the region of memory in coordination with the second client device is based on a memory-allocation data structure, wherein the memory-allocation data structure is shared with the second client device.

2 . The computer readable medium of claim 1 further comprising instructions executable to update the memory-allocation data structure to indicate that the portion has been allocated as part of the selection of the subset of the region of memory in coordination with the second client device.

3 . The computer readable medium of claim 2 , wherein the memory-allocation data structure is included in metadata associated with the region of memory.

4 . The computer readable medium of claim 1 , wherein the memory-allocation data structure is accessible from the first client device and the second client device via client-side memory access.

5 . The computer readable medium of claim 1 comprising instructions executable to access the memory-allocation data structure atomically as part of the selection of the subset of the region of memory in coordination with the second client device.

6 . The computer readable medium of claim 1 , wherein access to the memory-allocation data structure is coordinated based on one or more synchronization primitives.

7 . The computer readable medium of claim 1 , wherein the memory-allocation data structure is included in the region of memory of the memory appliance.

8 . The computer readable medium of claim 1 , wherein the memory-allocation data structure includes one or more indicators that indicate whether corresponding portions of the region are allocated.

9 . The computer readable medium of claim 1 , wherein selection of the subset of the region of memory in coordination with the second client device is based on message-passing between the first client and the second client.

10 . The computer readable medium of claim 1 , wherein the at least the subset of the region of memory mapped to the virtual address space includes all of the region of memory and wherein the at least the subset of the region of memory is mapped to the virtual address space before the request to allocate a portion of memory is received and/or before the subset of the region of memory is selected.

11 . The computer readable medium of claim 1 , wherein a second subset of the region of memory is allocable by the second client device in response to a request to allocate a second portion of memory.

12 . The computer readable medium of claim 1 further comprising instructions executable to initialize at least a second portion of external memory after the at least the second portion of external memory is mapped at the first client device.

13 . The computer readable medium of claim 1 further comprising instructions executable to free the subset of the region of memory in response to destruction of a corresponding application logic or to a determination that the corresponding application logic is destroyed.

14 . A method comprising:

receiving, at a first client device, a request to allocate a portion of memory, wherein a memory appliance and the first client device are connected by a network, and a region of memory of the memory appliance is allocated as external memory before receipt of the request to allocate the portion of memory at the first client device;

selecting, by a first client logic within the first client device and in coordination with a second client logic within a second client device on the network, a subset of the region of memory for the portion of memory allocated, wherein the portion of memory allocated is external to the first client device but is primary memory to the first client device, wherein by coordination with the second client device, the subset of the region of memory is selected without the second client device having selected the same subset of the region of memory for allocation as external memory for the second client device; and wherein the first client logic is otherwise distinct from the second client logic aside from coordinating with the second client logic in using the memory of the memory appliance; and

mapping, at the first client device, at least the subset of the region of memory to a virtual address space,

wherein selecting the subset of the region of memory in coordination with the second client device comprises accessing a memory-allocation data structure that is shared by the first client device and the second client device, the memory-allocation data structure indicating which portions of the region of memory are allocated by a memory user of the first client device.

15 . The method of claim 14 , wherein selecting the subset of the region of memory in coordination with the second client comprises identifying a free portion of the region of memory of the memory appliance from the memory-allocation data structure shared by the first client and the second client.

16 . The method of claim 14 further comprising migrating an application logic from the first client device to the second client device by mapping, at the second client, the portion of the region of memory of the memory appliance that is indicated by the memory-allocation data structure to be allocated to the application logic on the first client.

17 . The method of claim 16 , wherein migrating the application logic includes migrating the application logic from the first client device to the second client device without copying any data of the allocated portion of memory from the first client device to the second client device.

18 . The method of claim 16 further comprising determining the portion of the region of memory of the memory appliance allocated to the application logic on the first client by accessing the memory-allocation data structure from the second client device.

Assignments (2)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded May 7, 2026
From: KOVE IO, INC.
To: KOVE IP, LLC
Reel/Frame 074591/0184 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded May 7, 2026
From: STABRAWA, TIMOTHY A.; CORNELIUS, ZACHARY A.; OVERTON, JOHN; POLING, ANDREW S.; TAYLOR, JESSE I.
To: KOVE IO, INC.
Reel/Frame 074590/0946 →
Continuity (6)
Continuation 17125542 · Dec 17, 2020
Continuation 16395329 · Apr 26, 2019
Division 14854657 · Sep 15, 2015
Continuation In Part 14530908 · Nov 3, 2014
Provisional Application 62051144 · Sep 16, 2014
Related Publication 20220317903A1 · Oct 6, 2022
References Cited (116)
US 6049853A · Kingsbury et al. · 2000 [cited by applicant]
US 7756943B1 · Wong · 2010 [cited by applicant]
US 7925711B1 · Gopalan et al. · 2011 [cited by applicant]
US 8307154B2 · Stabrawa et al. · 2012 [cited by applicant]
US 8510524B1 · Corbett et al. · 2013 [cited by applicant]
US 8745276B2 · Bloch et al. · 2014 [cited by applicant]
US 8775755B2 · Magenheimer et al. · 2014 [cited by applicant]
US 8799914B1 · Metcalf · 2014 [cited by applicant]
US 8914458B2 · Raindel et al. · 2014 [cited by applicant]
US 9015426B2 · Stabrawa et al. · 2015 [cited by applicant]
US 9575889B2 · Chang et al. · 2017 [cited by applicant]
US 9619268B2 · Beveridge et al. · 2017 [cited by applicant]
US 10235064B1 · Natanzon et al. · 2019 [cited by applicant]
US 20030188114A1 · Lubbers et al. · 2003 [cited by applicant]
US 20030188218A1 · Lubbers et al. · 2003 [cited by applicant]
US 20040093389A1 · Mohammed et al. · 2004 [cited by applicant]
US 20040143562A1 · Chen et al. · 2004 [cited by applicant]
US 20040143718A1 · Chen et al. · 2004 [cited by applicant]
US 20040221102A1 · Watanabe · 2004 [cited by applicant]
US 20050044301A1 · Vasilevsky et al. · 2005 [cited by applicant]
US 20050204045A1 · Belkin et al. · 2005 [cited by applicant]
US 20050216552A1 · Fineberg et al. · 2005 [cited by applicant]
US 20060236063A1 · Hausauer et al. · 2006 [cited by applicant]
US 20080082593A1 · Komarov et al. · 2008 [cited by applicant]
US 20080104337A1 · VelurEunni · 2008 [cited by applicant]
US 20080235409A1 · Ryzhykh · 2008 [cited by applicant]
US 20090113425A1 · Ports et al. · 2009 [cited by applicant]
US 20100250867A1 · Bettger et al. · 2010 [cited by applicant]
US 20110264949A1 · Ikeuchi et al. · 2011 [cited by applicant]
US 20120079318A1 · Colgrove et al. · 2012 [cited by applicant]
US 20120221803A1 · Stabrawa et al. · 2012 [cited by applicant]
US 20120222052A1 · Magenheimer et al. · 2012 [cited by applicant]
US 20130185720A1 · Tuch et al. · 2013 [cited by applicant]
US 20130290546A1 · Samih et al. · 2013 [cited by applicant]
US 20140089451A1 · Eran et al. · 2014 [cited by applicant]
US 20140143368A1 · Anderson · 2014 [cited by applicant]
US 20140164545A1 · Davis et al. · 2014 [cited by applicant]
US 20150020071A1 · Phelan et al. · 2015 [cited by applicant]
US 20150293881A1 · Raikin et al. · 2015 [cited by applicant]
US 20150370494A1 · Wang · 2015 [cited by applicant]
US 20160019117A1 · Kumarasamy et al. · 2016 [cited by applicant]
US 20160077761A1 · Stabrawa et al. · 2016 [cited by applicant]
US 20160077966A1 · Stabrawa et al. · 2016 [cited by applicant]
US 20160077975A1 · Stabrawa et al. · 2016 [cited by applicant]
US 20170147227A1 · Stabrawa et al. · 2017 [cited by applicant]
US 20170147437A1 · Borlick et al. · 2017 [cited by applicant]
US 20170160979A1 · Golander et al. · 2017 [cited by applicant]
US 20170192865A1 · Pan · 2017 [cited by applicant]
US 20170285997A1 · Stabrawa et al. · 2017 [cited by applicant]
US 20170286702A1 · Cheung · 2017 [cited by applicant]
US 20180081559A1 · Stabrawa et al. · 2018 [cited by applicant]
WO WO2008132760A2 · 2008 [cited by applicant]
Liran Liss, “On Demand Paging for User-level Networking,” dated Sep. 30, 2013, pp. 1-25, Mellanox Technologies, Sunnyvale, CA. [cited by applicant]
Liran Liss, “On-Demand-Paging (ODP) Update,” dated Apr. 28, 2014, available at https:www.youtube.com/watch?v=KbrlsXQbHCw. [cited by applicant]
Brian Van Essen et al., “DI-MMAP—a scalable memory-map runtime for out-of-core data-intensive applications,” Cluster Computing dated Oct. 4, 2013, pp. 15-28, vol. 18, Issue 1, published by Springer, New York City, NY. [cited by applicant]
Hannes Mühleisen et al., “Peak Performance—Remote Memory Revisited,” dated Jun. 24, 2013, pp. 1-7, DaMoN' 13, New York, NY. [cited by applicant]
Kevin Te-Ming Lim, “Disaggregated Memory Architectures for Blade Servers,” dated 2010, pp. 1-148, published by The University of Michigan, Ann Arbor, MI. [cited by applicant]
Shuang Liang et al. “Swapping to Remote Memory over Infiniband: An Approach using a High Performance Network Block Device,” dated Sep. 2005, pp. 1-10, published by The Ohio State University, Columbus OH. [cited by applicant]
Shuang Liang et al. “Swapping to Remote Memory over Infiniband: An Approach using a High Performance Network Block Device,” dated Sep. 2005, pp. 1-30, Cluster Computing, 2005, IEEE International, Conference located in B… [cited by applicant]
Stephen Mathew Rumble, “Memory and Object Management in RAMCloud,” dated Mar. 2014, pp. 1-161, published by Stanford University, Stanford, CA. [cited by applicant]
Katnegermis, “rdma/rdma_common.c at master katnegermis/rdma,” GitHub, dated Apr. 27, 2014, pp. 1-4, available at http://githib.com/katnegermis/rdma/blob/master/rdma_common.c. [cited by applicant]
Katnegermis, rdma/rdma_server.c at master katnegermis/rdma, GitHub, dated May 6, 2014, pp. 1-7, available at http://github.com/katnegermis/rdma/blob/master/rdma_server.c. [cited by applicant]
Robert Russell, “Introduction to RDMA Programming”, dated Apr. 17, 2014, pp. 1-76, available at http://web.archive/org/web/20140417205540/http://www.cs.unh.edu/˜rdr/rdma-intro-module.ppt. [cited by applicant]
Anonymous, “Implementing calloc( ),” dated Apr. 12, 2010, pp. 1-3, published online by Lockless, available at URL http://web.archive.org/web/20100412183814/http://locklessinc.com/articles/calloc. [cited by applicant]
Konstantinos Karampogias, “An In-Memory RDMA-Based Architecture for the Hadoop Distributed Filesystem Master Thesis,” dated Aug. 21, 2012, pp. 1-49, published by IBM Research—Zurich, Rüschlikon, Switzerland, available a… [cited by applicant]
Anonymous, “Remoted direct memory access,” dated May 12, 2016, pp. 1-2, published online by Wikipedia at URL https://en.wikipedia.org/wiki/Remote_direct_memory_access. [cited by applicant]
Anonymous, “Copy-on-write,” dated Nov. 26, 2014, pp. 1-3, Wikipedia, published by Wikipedia at URL: https://en.wikipedia.org/w/index.php?title=Copy-on-write&oldid=635508783. [cited by applicant]
Chao Wang et al., “NVMalloc: Exposing an Aggregate SSD as a Memory Partition in Extreme-Scale Machines,” dated 2012, pp. 957-968, Parallel & Distributed Processing Symposium (IPDPS), 2012 IEEE 26th International, Shangh… [cited by applicant]
“Hugetlbpage”, dated Sep. 13, 2015, pp. 1-6, The Linux Kernel Archives, available at URL https://www.kernel.org/doc/Documentation/vm/hugetlbpage.txt. [cited by applicant]
Iulian Moraru, “nvmalloc”, dated Feb. 27, 2013, p. 1, published online by GitHub at URL https://github.com/efficient/nvram/tree/master/nvmalloc. [cited by applicant]
Christopher Dall, et al., “KVM/ARM: The Design and Implementation of the Linux ARM Hypervisor,” dated Mar. 1, 2014, pp. 333-347, ASPLOS '14, Mar. 1-4, 2014, Salt Lake City, UT, Copyright is held by the owner/author(s). … [cited by applicant]
Avi Kivity et al., “kvm: the Linux Virtual Machine Monitor,” dated Jun. 27, 2007, pp. 225-232, Proceedings of the Linux Symposium, vol. 1, published online by Kernel, Ottawa, Ontario, Canada. [cited by applicant]
Josh Triplett, “Using the KVM API,” dated Sep. 29, 2015, pp. 1-12, published by LWN.net at URL https://lwn.net/Articles/658511. [cited by applicant]
Jes Sorensen, “KVM Live Snapshot support,” dated Jun. 1, 2011, pp. 1-22, published by RedHat, LinuxCon, Japan. [cited by applicant]
“Warning: The xenpaging code is new and not fully debugged . . . ”, p. 1, Retrieved from the Internet on Jan. 27, 2016, at URL: https://xenbits.xen.org/docs/4.3-testing/.misc/xenpaging.txt. [cited by applicant]
Brian Van Essen, “tagged_page.c” (source code file is part of DI-MMAP, Version 1.0), dated 2012, pp. 1-13, Lawrence Livermore National Security, LLC, retrieved from URL https://bitbucket.org/vanessen/di-mmap/src. [cited by applicant]
Brian Van Essen, “Snippet of ‘tagged_page.c’” (source code file is part of DI-MMAP, Version 1.0), dated 2012, p. 1, Lawrence Livermore National Security, LLC, retrieved from URL https://bitbucket.org/vanessen/di-mmap/sr… [cited by applicant]
Brian Van Essen, “helpers.c” (source code file is part of DI-MMAP, Version 1.0), dated 2012, pp. 1-19, Lawrence Livermore National Security, LLC, retrieved from URL https://bitbucket.org/vanessen/di-mmap/src. [cited by applicant]
Brian Van Essen, “Snippet of ‘helpers.c’”(source code file is part of DI-MMAP, Version 1.0), dated 2012, p. 1, Lawrence Livermore National Security, LLC, retrieved from URL https://bitbucket.org/vanessen/di-mmap/src. [cited by applicant]
Anonymous, “Memory Resource Controller” (memory cgroup documentation—Kernel.org), dated Jan. 25, 2016, pp. 123, published by Kernel.org at URL https://www.kernel.org/doc/Documentation/cgroup-v1/memory.txt. [cited by applicant]
Ahmed Al-Moayed et al., “Quality of Service Attributes in Web Services,” dated 2010, pp. 367-372, 2010 Fifth International Conference on Software Engineering Advances, © 2010 IEEE. [cited by applicant]
Aleksandar Dragojević et al., FaRM: Fast Remote Memory, Proceedings of the 11th USENIX Symposium on Networked Systems Design and Implementation (NSDI '14), ISBN 978-1-931971-09-6, dated Apr. 2, 2014, pp. 1-15, published… [cited by applicant]
International Search Report, issued in International Application No. PCT/US2015/050170, dated Dec. 8, 2015, pp. 1-14, European Patent Office, Rijswijk, Netherlands. [cited by applicant]
Partial International Search Report, issued in International Application No. PCT/US2015/050177, dated Dec. 17, 2015, pp. 1-8, European Patent Office, Rijswijk, Netherlands. [cited by applicant]
International Search Report, issued in International Application No. PCT/US2015/050177, dated May 4, 2016, pp. 1-28, European Patent Office, Rijswijk, Netherlands. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 14/530,908, dated Apr. 8, 2016, pp. 1-21, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 14/554,655, dated Apr. 6, 2016, pp. 1-22, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 14/854,657, dated Oct. 23, 2017, pp. 1-12, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Final Office Action, issued in U.S. Appl. No. 14/530,908, dated Jul. 5, 2016, pp. 1-22, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Extended European Search Report, issued in European Patent Application No. 16162354.1, dated Jul. 28, 2016, pp. 1-11, European Patent Office, Munich, Germany. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 15/621,537, dated Jun. 29, 2017, pp. 1-14, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 14/554,655, dated Jan. 11, 2017, pp. 1-8, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 15/800,891, dated Apr. 2, 2018, pp. 1-14, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Final Office Action, issued in U.S. Appl. No. 14/854,657, dated Apr. 5, 2018, pp. 1-13, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 15/424,395, dated Apr. 2, 2018, pp. 1-17, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Final Office Action, issued in U.S. Appl. No. 15/424,395, dated Dec. 18, 2018, pp. 1-11, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Final Office Action, issued in U.S. Appl. No. 14/854,657, dated Aug. 24, 2018, pp. 1-6, United States Patent and Trademark Office, Alexandria, VA. [cited by applicant]
European Office Action, issued in European Application No. 15782105.9, dated Nov. 29, 2018, pp. 1-7, European Patent Office Rijswijk, Netherlands. [cited by applicant]
European Office Action, issued in European Application No. 15 770 761.3, dated Nov. 29, 2018, pp. 1-7, European Patent Office, Rijswijk, Netherlands. [cited by applicant]
Kevin Lim et al., “Disaggregated Memory for Expansion and Sharing in Blade Servers,” International Symposium on Computer Architecture, dated Jun. 20, 2009, pp. 1-2, published by ACM, Austin, TX. [cited by applicant]
Final Office Action, issued in U.S. Appl. No. 15/800,891, dated Jul. 9, 2018, pp. 1-16, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 16/050,974, dated Dec. 10, 2019, pp. 1-17, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 16/050,974, dated Oct. 27, 2020, pp. 1-11, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Extended European Search Report, issued in European Patent Application No. 20207722.8, dated Dec. 4, 2020, European Patent Office, Munich, Germany. [cited by applicant]
Andreas Hollmann et al., “Invasive Computing: An Application Assisted Resource Management Approach,” Multicore Software Engineering, Performance, and Tools, dated May 31, 2012, pp. 82-85, published by Springer-Verlag, B… [cited by applicant]
Emmanuel Cecchet, “Memory Mapped Networks: a new deal for Distributed Shared Memories? The SciFS experience,” ISBN 978-0-7695-1745-2, Cluster Computing, 2002, dated Sep. 23, 2002, pp. 231-238, 2002 IEEE International Co… [cited by applicant]
Non-Final Office Action, issued in U.S. Appl. No. 16/395,329, dated Dec. 22, 2020, pp. 1-19, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Yukun Liu et al., “The Development Tutorial via UNIX Kernel Services,” UNIX Operating System, online SBN 978-3-642-20432-6, dated 2011, pp. 1-368, published online by Springer Berlin Heidelberg at URL https://doi.org/10… [cited by applicant]
Non-Final Office Action, pp. 1-19, dated Dec. 22, 2020, issued in U.S. Appl. No. 16/395,329, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Yukun Liu et al., “The Development Tutorial via UNIX Kernel Services,” pp. 1-368, dated 2011, UNIX Operating System, online SBN 978-3-642-20432-6, published online by Springer Berlin Heidelberg at URL https://doi.org/10… [cited by applicant]
Anirudh Badam and Vivek S. Pai, “SSDAlloc: Hybrid SSD/RAM Memory Management Made Easy”, pp. 1-16, dated 2011, www.microsoft.com, available at https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/Badam.pdf. [cited by applicant]
European Office Action, dated Jan. 26, 2022, pp. 1-7, issued in European Application 20192905.6, European Patent Office, Munich, Germany. [cited by applicant]
Final Office Action, dated Feb. 18, 2022, pp. 1-21, issued in U.S. Appl. No. 16/530,651, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Notice of Allowance, dated Mar. 1, 2022, pp. 1-11, issued in U.S. Appl. No. 16/395,329, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Notice of Allowance, issued in U.S. Appl. No. 16/395,329, dated Mar. 1, 2022, pp. 1-11, U.S. Patent and Trademark Office, Alexandria, VA. [cited by applicant]
Giloi et al., A Distributed Implementation of Shared Virtual Memory With Strong and Weak Coherence, dated 1991, pp. 1-9, GMD Research Center (Berlin, Germany). [cited by applicant]