International Business Machines Corporation

Hewlett Packard Enterprise Development LP

Prashant Anant Paranjape

Nikhil Hegde

David Richard Marquardt

Seth Daniel Jennings

Roocha K Pandya

Vasu Vallabhaneni

Yogesh L Hegde

Chandan Chopra

Dwip N Banerjee

Swati Rallapalli

Patrick Tam Vo

Poornima Srinivas Patil

Paul E Stolle

Sandhiya Krishnasamy

Aspects pertaining to a crash dump file are described. In an example, the crash dump file may include one or more sections within which crash-related data is stored. A content file is also described which records valid parameters, wherein the valid section parameters pertain to corresponding sections within the crash dump file.

Sections in crash dump files

A system, and computer usable program product for addressing a workload partition are provided in the illustrative embodiments. A packet of data directed to the WPAR and including a combined address is received at a software stack. A determination is made whether the combined address includes an identifier that is reserved. If the identifier is not reserved, the identifier is used to identify the WPAR in the data processing system and the packet is sent to the WPAR. The software stack is an application configured to enable data communication using a protocol of the packet, and is modified to detect, recognize, and process the combined address for routing the packet to the WPAR. The combined address may be logically formed using an identifier associated with the data processing system and an identifier associated with the WPAR, which may be located in different portions of the packet.

Addressing a workload partition

A method includes receiving, into a target machine communicatively coupled to a first subnetwork, data for migration of a virtual partition from a source machine communicatively coupled to a second subnetwork. The virtual partition has a source network address defined by network addressing of the second subnetwork. The method includes initiating restart of the virtual partition on the target machine, wherein the initiating comprises determining a socket connection providing network communication between the virtual partition and a host and transmitting, through a network tunnel communication from the target machine to the host, a communication providing a target network address that is assigned to the virtual partition after the virtual partition is migrated to the target machine. The method includes routing a packet, from a process executing in the virtual partition of the target machine, to the host based on the virtual partition having the target network address.

Migration of logical partitions between two devices

A method for addressing a workload partition is provided in the illustrative embodiments. A packet of data directed to the WPAR and including a combined address is received at a software stack. A determination is made whether the combined address includes an identifier that is reserved. If the identifier is not reserved, the identifier is used to identify the WPAR in the data processing system and the packet is sent to the WPAR. The software stack is an application configured to enable data communication using a protocol of the packet, and is modified to detect, recognize, and process the combined address for routing the packet to the WPAR. The combined address may be logically formed using an identifier associated with the data processing system and an identifier associated with the WPAR, which may be located in different portions of the packet.

Blind attacks on a protocol connection, such as a TCP connection, are prevented by inserting checksums computed during protocol connection establishment handshake into data sent through the connection and invalidating data sent through the connection that lacks the protocol setup information checksums. Reset attacks are prevented by invalidating reset requests unless a master checksum computed from the protocol setup information checksums is included with the reset request. Checksums computed from protocol setup information have improved robustness by including a random number with the protocol setup information.

Securing network communications from blind attacks with checksum comparisons

Mechanisms are provided for resolving a domain name. The mechanisms send a request, from a resolver executing on the data processing system, to a name server requesting a resolution of a domain name into an address. A response to the request is received that includes a local address. The local address has a scope that is local to a site and is not global in scope. A determination is made as to whether the local address corresponds to a domain associated with the resolver. Communications are then performed with a device corresponding to the domain name using the local address in response to a determination that the local address corresponds to a domain associated with the resolver.

Using unique local unicast addresses in a global domain name server

A method identifies a network socket on a machine by assigning a non-IP address socket anchor and a port number to the network socket. A socket anchor may comprise a globally unique number. In some embodiments, a socket anchor comprises a number having a preselected number of bits. The number having the preselected number of bits may comprise a selected number associated with the machine concatenated with a number of randomly generated bits. The number of bits of the selected number plus the number of randomly generated bits is equal to the preselected number of bits. In some embodiments, the selected number associated with the machine is a Media Access Control (MAC) address of a network interface associated with the machine.

Transport layer connection

A new instance of a connection that is in the wait state can be established while mitigating connection identifier collisions. Based on a connection identifier in a connection request received at a server from a client, it is determined that a previous instance of a connection represented by the connection identifier is in a wait state. An acknowledgement message comprising wait state parameters associated with the connection identifier and original connection parameters associated with the previous instance of the connection in the wait state is provided from the server to the client. In response, the server receives a compliance message comprising updated wait state parameters and updated connection parameters from the client. The new instance of the connection represented by the connection identifier that previously represented the previous instance of the connection is established based, at least in part, on the updated wait state parameters and the updated connection parameters.

Mitigating connection identifier collisions in a communication network

A method includes receiving a command to migrate a workload partition from a source machine to a target machine. The source machine and the target machine are communicatively coupled to a first subnetwork and a second subnetwork, respectively. The workload partition has a source network address defined by network addressing of the first subnetwork. The method also includes determining a socket connection providing network communication between the workload partition and a host. The method includes notifying the host that the workload partition is migrating from the source machine to the target machine, wherein in response to the notifying the host is on alert to receive, through a network tunnel communication, an identification of a target network address that is assigned to the workload partition after the workload partition is migrated to the target machine. The workload partition is migrated from the source machine to the target machine.

A system identifies a network socket on a machine by assigning a non-IP address socket anchor and a port number to the network socket. A socket anchor may comprise a globally unique number. In some embodiments, a socket anchor comprises a number having a preselected number of bits. The number having the preselected number of bits may comprise a selected number associated with the machine concatenated with a number of randomly generated bits. The number of bits of the selected number plus the number of randomly generated bits is equal to the preselected number of bits. In some embodiments, the selected number associated with the machine is a Media Access Control (MAC) address of a network interface associated with the machine.

Mechanisms are provided for determining a path maximum transmission unit of a path between a source node and a destination node of the path. These mechanisms transmit an echo request packet from the source node to the destination node along the path. The mechanisms further determine, by at least one intermediate node along the path, whether a size of the echo request packet is greater than a maximum transmission unit (MTU) of a next link of the path from the intermediate node to a next node along the path. Moreover, the mechanisms record, by the at least one intermediate node, the MTU of the next link in the echo request packet in response to the determination that the size of the echo request packet is greater than the MTU of the next link.

Mechanisms for discovering path maximum transmission unit

Mechanisms are provided for assigning addresses for use in communications between computing devices. With these mechanisms, a centralized domain name registrar server assigns a unique global identifier to a domain of a site such that the unique global identifier uniquely identifies the domain. The unique global identifier is received in a data processing system from the centralized domain name registrar server. The site is divided into a plurality of sub-domains and a unique sub-domain identifier is assigned to each sub-domain in the site such that the unique sub-domain identifier uniquely identifies a corresponding sub-domain in the site. The global identifier for the domain of the site and the sub-domain identifiers of the sub-domains in the plurality of sub-domains are provided by the data processing system to routers of the site for use in associating addresses with devices of the site.

Using unique local unicast addresses in a global domain name server by providing a centralized registry

A method, system, and computer usable program product for securing a data communication against attacks are provided in the illustrative embodiments. A segment in the data communication is received at a first application executing in a first data processing system. The segment is formed according to a data communication protocol and includes an option. The option includes a current clue and a next clue. The current clue is compared with a saved next clue, the saved next clue being a next clue in a previous segment. The segment is accepted as being a valid segment in the data communication if the current clue matches the saved next clue. A part of the segment is sent to a consumer application.

Securing a communication protocol against attacks

Disclosed is a computer implemented method and computer program product for transmitting a resource record to a requesting computer. An authoritative domain name server receives a DNS query from a requesting computer at a name server. The authoritative domain name server looks up the resource record based on the DNS query, wherein the resource record is associated with an epochal time and a time to live. The authoritative domain name server transmits the resource record response based on the epochal time.

Growing community of inventors

Bengaluru, India

Prashant Anant Paranjape