Patents by Inventor Dimitry Melts
Dimitry Melts has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11397703Abstract: Embodiments described herein provide a method for accessing a host memory through non-volatile memory over fabric bridging with direct target access. A first memory access command encapsulated in a first network packet is received at a memory interface unit and from a remote direct memory access (RDMA) interface and via a network fabric. The first memory access command is compliant with a first non-volatile memory interface protocol and the first network packet is compliant with a second non-volatile memory interface protocol. The first network packet is unwrapped to obtain the first memory access command. The first memory access command is stored in a work queue using address bits of the work queue as a pre-set index of the first memory access command. The first memory access command is sent from the work queue based on the pre-set index to activate a first target storage device.Type: GrantFiled: September 4, 2020Date of Patent: July 26, 2022Assignee: Marvell Asia Pte, Ltd.Inventors: Eitan Joshua, Scott Furey, Dimitry Melts, Noam Mizrahi
-
Publication number: 20200401551Abstract: Embodiments described herein provide a method for accessing a host memory through non-volatile memory over fabric bridging with direct target access. A first memory access command encapsulated in a first network packet is received at a memory interface unit and from a remote direct memory access (RDMA) interface and via a network fabric. The first memory access command is compliant with a first non-volatile memory interface protocol and the first network packet is compliant with a second non-volatile memory interface protocol. The first network packet is unwrapped to obtain the first memory access command. The first memory access command is stored in a work queue using address bits of the work queue as a pre-set index of the first memory access command. The first memory access command is sent from the work queue based on the pre-set index to activate a first target storage device.Type: ApplicationFiled: September 4, 2020Publication date: December 24, 2020Inventors: Eitan Joshua, Scott Furey, Dimitry Melts, Noam Mizrahi
-
Patent number: 10769098Abstract: Embodiments described herein provide a method for accessing a host memory through non-volatile memory over fabric bridging with direct target access. A first memory access command encapsulated in a first network packet is received at a memory interface unit and from a remote direct memory access (RDMA) interface and via a network fabric. The first memory access command is compliant with a first non-volatile memory interface protocol and the first network packet is compliant with a second non-volatile memory interface protocol. The first network packet is unwrapped to obtain the first memory access command. The first memory access command is stored in a work queue using address bits of the work queue as a pre-set index of the first memory access command. The first memory access command is sent from the work queue based on the pre-set index to activate a first target storage device.Type: GrantFiled: April 4, 2017Date of Patent: September 8, 2020Assignee: Marvell Asia Pte, Ltd.Inventors: Eitan Joshua, Scott Furey, Dimitry Melts, Noam Mizrahi
-
Patent number: 10416887Abstract: A hybrid storage device includes at least a first storage device operating under a first storage interface protocol, at least a second storage device operating under a second storage interface protocol, and a drive controller interface. The drive controller interface has a front-end for connecting to a host via a host interface protocol, a back-end for connecting to the first storage device via the first storage interface protocol, and to the second storage device via the second storage interface protocol, and a respective translation module for translating between the host interface protocol and a respective one of the first and second storage interface protocols. The hybrid storage device may be included in a hybrid storage system with a host processor. The host interface protocol may be the same as one of the first and second storage interface protocols, such as NVMe. The protocols may be implemented in hardware or software.Type: GrantFiled: May 17, 2017Date of Patent: September 17, 2019Assignee: Marvell International Ltd.Inventors: Dimitry Melts, Scott Furey
-
Patent number: 10243865Abstract: A network device includes (i) a software forwarding engine, and (ii) a hardware forwarding engine, wherein the software forwarding engine is implemented using a processor executing machine readable instructions. The network device analyzes a header of a received packet to determine i) whether the received packet belongs to any flows of packets already known to the network device, and ii) a packet type of the received packet. The network device selects one of the software forwarding engine or the hardware forwarding engine to process the received packet based on i) whether the received packet belongs to any flows of packets already known to the network device, and ii) the determined packet type, including selecting the software forwarding engine when it is determined that the received packet does not belong to any flow of packets already known to the network device.Type: GrantFiled: March 9, 2017Date of Patent: March 26, 2019Assignee: Marvell Israel (M.I.S.L) Ltd.Inventors: Erez Izenberg, Alon Pais, Ruven Torok, Dimitry Melts, Yuval Caduri, Dmitri Epshtein
-
Patent number: 10056147Abstract: A data storage device includes a first data storage medium having a first capacity and a first speed, a second data storage medium having a second capacity and a second speed, and a device controller for interfacing between the data storage device and a host system. The second capacity is greater than the first capacity and the second speed is slower than the first speed. The device controller presents the data storage device to the host system as having a device capacity at least equal to the second capacity and a device speed at least equal to the first speed. The first data storage medium may be a solid-state drive while the second data storage medium is a hard disk drive. The device controller may be a solid-state drive controller, or a hard disk drive controller that may accept at least one solid-state drive command, such as a TRIM command.Type: GrantFiled: May 24, 2017Date of Patent: August 21, 2018Assignee: Marvell International Ltd.Inventors: Pantelis Alexopoulos, Dimitry Melts, Christophe Therene, Shinichiro Kuno
-
Publication number: 20170286363Abstract: Embodiments described herein provide a method for accessing a host memory through non-volatile memory over fabric bridging with direct target access. A first memory access command encapsulated in a first network packet is received at a memory interface unit and from a remote direct memory access (RDMA) interface and via a network fabric. The first memory access command is compliant with a first non-volatile memory interface protocol and the first network packet is compliant with a second non-volatile memory interface protocol. The first network packet is unwrapped to obtain the first memory access command. The first memory access command is stored in a work queue using address bits of the work queue as a pre-set index of the first memory access command. The first memory access command is sent from the work queue based on the pre-set index to activate a first target storage device.Type: ApplicationFiled: April 4, 2017Publication date: October 5, 2017Inventors: Eitan Joshua, Scott Furey, Dimitry Melts, Noam Mizrahi
-
Publication number: 20170180264Abstract: A network device includes (i) a software forwarding engine, and (ii) a hardware forwarding engine, wherein the software forwarding engine is implemented using a processor executing machine readable instructions. The network device analyzes a header of a received packet to determine i) whether the received packet belongs to any flows of packets already known to the network device, and ii) a packet type of the received packet. The network device selects one of the software forwarding engine or the hardware forwarding engine to process the received packet based on i) whether the received packet belongs to any flows of packets already known to the network device, and ii) the determined packet type, including selecting the software forwarding engine when it is determined that the received packet does not belong to any flow of packets already known to the network device.Type: ApplicationFiled: March 9, 2017Publication date: June 22, 2017Inventors: Erez IZENBERG, Alon PAIS, Ruven TOROK, Dimitry MELTS, Yuval CADURI, Dmitri EPSHTEIN
-
Patent number: 9614755Abstract: A forwarding system comprises a identification engine, a hardware forwarding engine configured to process an ingressing packet, a software forwarding engine configured to process the ingressing packet, and a selection engine. The selection engine is configured to select one of the hardware forwarding engine or the software forwarding engine to process the ingressing packet. The selection is based on at least one of an indication of resource availability or a classification of the ingressing packet based on a priority of a flow as determined by the identification engine. In some embodiments, the selection engine selects different forwarding engines to process different packets of a same flow based on changes in resource availability or classification of the ingressing packet.Type: GrantFiled: September 29, 2014Date of Patent: April 4, 2017Assignee: Marvell Israel (M.I.S.L) Ltd.Inventors: Erez Izenberg, Alon Pais, Ruvel Torok, Dimitry Melts, Yuval Caduri, Dimitri Epshtein
-
Patent number: 9344215Abstract: Systems and methods are provided for customer premises equipment (CPE) on a passive optical network (PON). A system includes a packet processor having at least an active mode and a sleep mode, the packet processor configured to processes streams of data packets received in a data plane from an optical line terminal (OLT) through the PON when in an active mode and to enter the sleep mode when not receiving data packets in the data plane. A system further includes a micro-controller, separate from the packet processor, configured to receive from an OLT operation and management (OAM) messages that are transmitted in a control plane, and to process the OAM messages by, selectively transmitting to a central office, without waking up the packet processor, an acknowledgement message, or waking up the packet processor to receive data packets in the data plane.Type: GrantFiled: September 17, 2013Date of Patent: May 17, 2016Assignee: MARVELL WORLD TRADE LTD.Inventors: Dimitry Melts, Roman Kleinerman, Amit Avivi
-
Patent number: 9338530Abstract: Methods and systems for implementing versatile optical terminals that detect optical transmission protocols and subsequently adapt to the correct protocol are disclosed. In an embodiment, an interface device for providing an interface for a first network with a passive optical network (PON) is disclosed. The interface device includes a protocol detection circuit for determining whether optical communication signals received from the PON conform to a first optical communication protocol, and a switchover control circuit that reconfigures the interface device to work with a second optical communication protocol when the received optical communication signals do not conform to the first optical communication protocol.Type: GrantFiled: January 22, 2014Date of Patent: May 10, 2016Assignee: Marvell Israel (M.I.S.L) Ltd.Inventors: Erez Izenberg, Oren Ben-Hayune, Erez Amit, Dimitry Melts, Arie Elias
-
Patent number: 9307302Abstract: A Passive Optical Network (PON) Switch which breaks down and regenerates a point to multipoint optical communication signals that are compliant with a PON protocol between an Optical Line Terminal (OLT) and an Optical Network Unit (ONU) by performing a conversion between optical communication signals compliant with PON protocol and data units compliant with Ethernet protocol.Type: GrantFiled: November 11, 2013Date of Patent: April 5, 2016Assignee: Marvell World Trade Ltd.Inventors: Israel Ben-Shahar, Dimitry Melts, Roman Kleinerman
-
Patent number: 9036993Abstract: Embodiments of the present disclosure provide methods for allocating bandwidth to a plurality of traffic containers of a passive optical network. The method comprises receiving upstream data from a plurality of traffic containers of the passive optical network and passing the upstream data to a traffic manager. The method further comprises dynamically changing the allocated bandwidth based at least in part on the amount of the upstream data stored in one or more queues of the traffic manager.Type: GrantFiled: February 12, 2013Date of Patent: May 19, 2015Assignee: Marvell World Trade Ltd.Inventors: Dimitry Melts, Tsahi Daniel
-
Publication number: 20150016451Abstract: A forwarding system comprises a identification engine, a hardware forwarding engine configured to process an ingressing packet, a software forwarding engine configured to process the ingressing packet, and a selection engine. The selection engine is configured to select one of the hardware forwarding engine or the software forwarding engine to process the ingressing packet. The selection is based on at least one of an indication of resource availability or a classification of the ingressing packet based on a priority of a flow as determined by the identification engine. In some embodiments, the selection engine selects different forwarding engines to process different packets of a same flow based on changes in resource availability or classification of the ingressing packet.Type: ApplicationFiled: September 29, 2014Publication date: January 15, 2015Inventors: Erez Izenberg, Alon Pais, Ruvel Torok, Dimitry Melts, Yuval Caduri, Dimitri Epshtein
-
Patent number: 8924810Abstract: A data unit is received, wherein the data unit includes a primary information data structure and a primary redundancy data structure. A secondary error correction operation is performed on one or more selected fields within the primary information data structure. After performing the secondary error correction operation, a primary error correction operation is performed on the data unit using the primary redundancy data structure, the primary error correction operation separate from the secondary error correction operation.Type: GrantFiled: December 16, 2013Date of Patent: December 30, 2014Inventors: Erez Izenberg, Oren Shafrir, Erez Amit, Dimitry Melts
-
Patent number: 8848715Abstract: A forwarding system comprises a identification engine, a hardware forwarding engine configured to process an ingressing packet, a software forwarding engine configured to process the ingressing packet, and a selection engine. The selection engine is configured to select one of the hardware forwarding engine or the software forwarding engine to process the ingressing packet. The selection is based on at least one of an indication of resource availability or a classification of the ingressing packet based on a priority of a flow as determined by the identification engine. In some embodiments, the selection engine selects different forwarding engines to process different packets of a same flow based on changes in resource availability or classification of the ingressing packet.Type: GrantFiled: March 1, 2011Date of Patent: September 30, 2014Assignee: Marvell Israel (M.I.S.L) Ltd.Inventors: Erez Izenberg, Alon Pais, Ruven Torok, Dimitry Melts, Yuval Caduri, Dmitri Epshtein
-
Patent number: 8839405Abstract: A physical layer device includes memory, a memory control module, and a physical layer module. The memory control module is configured to control access to the memory. The physical layer module is configured to store packets in the memory via the memory control module. The physical layer module includes an interface configured to receive the packets from a network device via a network and an interface bus. The interface bus includes at least one of a control module and a regular expression module. The at least one of the control module and the regular expression module is configured to inspect the packets to determine a security level of the packets. A network interface is configured to, based on the security level, provide the packets to a device separate from the physical layer device.Type: GrantFiled: August 10, 2012Date of Patent: September 16, 2014Assignee: Marvell World Trade Ltd.Inventors: Sehat Sutardja, Tsahi Daniel, Dimitry Melts
-
Publication number: 20140133853Abstract: A Passive Optical Network (PON) Switch which breaks down and regenerates a point to multipoint optical communication signals that are compliant with a PON protocol between an Optical Line Terminal (OLT) and an Optical Network Unit (ONU) by performing a conversion between optical communication signals compliant with PON protocol and data units compliant with Ethernet protocol.Type: ApplicationFiled: November 11, 2013Publication date: May 15, 2014Applicants: Marvell World Trade Ltd.Inventors: Israel BEN-SHAHAR, Dimitry MELTS, Roman KLEINERMAN
-
Publication number: 20140133854Abstract: Methods and systems for implementing versatile optical terminals that detect optical transmission protocols and subsequently adapt to the correct protocol are disclosed. In an embodiment, an interface device for providing an interface for a first network with a passive optical network (PON) is disclosed. The interface device includes a protocol detection circuit for determining whether optical communication signals received from the PON conform to a first optical communication protocol, and a switchover control circuit that reconfigures the interface device to work with a second optical communication protocol when the received optical communication signals do not conform to the first optical communication protocol.Type: ApplicationFiled: January 22, 2014Publication date: May 15, 2014Applicant: Marvell International Ltd.Inventors: Erez IZENBERG, Oren BEN-HAYUNE, Erez AMIT, Dimitry MELTS, Arie ELIAS
-
Publication number: 20140105603Abstract: Systems and methods are provided for customer premises equipment (CPE) on a passive optical network (PON). A system includes a packet processor having at least an active mode and a sleep mode, the packet processor configured to processes streams of data packets received in a data plane from an optical line terminal (OLT) through the PON when in an active mode and to enter the sleep mode when not receiving data packets in the data plane. A system further includes a micro-controller, separate from the packet processor, configured to receive from an OLT operation and management (OAM) messages that are transmitted in a control plane, and to process the OAM messages by, selectively transmitting to a central office, without waking up the packet processor, an acknowledgement message, or waking up the packet processor to receive data packets in the data plane.Type: ApplicationFiled: September 17, 2013Publication date: April 17, 2014Applicants: Marvell Israel (M.I.S.L) Ltd., Marvell International Ltd.Inventors: Dimitry Melts, Roman Kleinerman, Amit Avivi