International Business Machines Corporation

Joseph H Allen

David J Hoeweler

Gari Singh

Meeta Yadav

Moshe M E Matsa

Dinakaran B Joseph

Jeffrey M Peters

John Adam Shriver

Heather D Achilles

Eric David Perkins

David Zachary Maze

Content of a publish-subscribe topic tree is provided. This includes receiving a path for a requested topic. The path specifies topics leading from a highest to a lowest level topic in the path. Content in the topic tree is retrieved for matching topics that match the lowest level topic in the path independently of whether the matching topics are on the path in the tree.

Technology for providing content of a publish-subscribe topic tree

Embodiments of the invention provide for the optimization of utilization of non-volatile memory in message queuing. In an embodiment of the invention, a method for optimizing utilization of non-volatile memory in message queuing includes receiving a new message in a message queuing system implemented in a host computing system. The method also includes storing the new message as a master message in non-volatile memory of the host computing system. The method yet further includes subsequently receiving different messages that each share redundant information with the master message. The method even yet further includes delta encoding each of the different messages and storing the delta encoded different messages in the non-volatile memory. Finally, the method includes deleting the master message from the non-volatile memory only once each of the different messages and the master message have been acknowledged by at least one consumer subscribing to the message queuing system.

Optimization of non-volatile memory in message queuing

Embodiments of the invention provide for the optimization of utilization of non-volatile memory in message queuing. In an embodiment of the invention, a method for optimizing utilization of non-volatile memory in message queuing includes receiving a new message in a message queueing system implemented in a host computing system. The method also includes storing the new message as a master message in non-volatile memory of the host computing system. The method yet further includes subsequently receiving different messages that each share redundant information with the master message. The method even yet further includes delta encoding each of the different messages and storing the delta encoded different messages in the non-volatile memory. Finally, the method includes deleting the master message from the non-volatile memory only once each of the different messages and the master message have been acknowledged by at least one consumer subscribing to the message queuing system.

Provided are techniques for a garbage-collected interned string table. Reference objects that each reference strings in the interned string table are created. The reference objects and the strings referenced by the reference objects in the interned string table are marked with a current generation. Garbage collection is performed by: advancing the current generation to a next generation; in response to the reference objects in a previous generation from the current generation advancing to an acquired references state, promoting the reference objects and the strings referenced by the reference objects in the previous generation to the current generation; and deleting the strings that are marked with the previous generation in the interned string table.

Garbage collection of interned strings

A memory allocation message for each primary memory allocation in a primary memory made by a primary processor is received at a hardware memory management module. A representation of each primary memory allocation is allocated within a second memory in response to each memory allocation message. A determination is made, based upon the allocated representations of each primary memory allocation within the second memory, to free a primary memory allocation in the primary memory. A memory free message is sent to the primary processor instructing the primary processor to free the primary memory allocation in the primary memory.

Hardware off-load memory garbage collection acceleration

A computer implemented method for writing to a software bound ring buffer. A network adapter may determine that data is available to write to the software bound ring buffer. The network adapter determines that a read index is not equal to a write index, responsive to a determination that data is available to write to the software bound ring buffer. The network adapter writes the data to memory referenced by the hardware write index, wherein memory referenced by the write index is offset according to an offset, and the memory contents comprise a data portion and a valid bit. The network adapter writes an epoch value of the write index to the valid bit. The network adapter increments the write index, responsive to writing the data to memory referenced by the write index. Further disclosed is method to access a hardware bound ring buffer.

Control flow in a ring buffer

A memory allocation message for each primary memory allocation in a primary memory made by a primary processor is received at a hardware memory management module, including an indication of whether a finalizer routine is associated with each primary memory allocation. A representation of each primary memory allocation is allocated within a second memory in response to each memory allocation message, including the indication of whether there is the associated finalizer routine. A determination is made, based upon the allocated representations of each primary memory allocation within the second memory, to free a primary memory allocation in the primary memory. A call object finalizer message is sent to the primary processor instructing the primary processor to call the finalizer routine associated with the primary memory allocation in the primary memory in response to determining that the primary memory allocation has the associated finalizer routine.

Hardware off-load garbage collection acceleration for languages with finalizers

Ring buffer

A method, computer program product, and system are disclosed for supporting variable data structures and constant data structures in a hardware-based XML PPE. The method may include inserting one or more constant Type-Length-Attributes (TLAs) into an output stream. The constant TLAs are inserted from one or more constant node-sets referenced from a constant table. The method may also include creating one or more variable node-sets. Each variable node-set includes one or more variable TLAs (TLAs saved from a base XML document) and is mapped to a variable table. The method may also include inserting one or more variable TLAs into the output stream inserted from one or more variable node-sets referenced from the variable table.

Growing community of inventors

Belmont, MA, United States of America

Joseph H Allen