Amazon Technologies, Inc.

Colin John Whittaker

Andrew Bruce Dickinson

Alan Michael Judge

Frederick David Sinn

Bradley D Roberts

Justin Oliver Pietsch

Rachit Chawla

Richendra Khanna

Eric Paul Wei

Mark Noel Kelly

Richard Hazen Galliher, Iii

Albert Park Niemoller

Joseph L Craig

Leonard Thomas Tracy

John William Evans

Benjamin Tobler

Alaa Adel Mahdi Hayder

Stephen Callaghan

Ke Zhang

Arijit Ganguly

Christopher J Lefelhocz

Nigel Martin McGee

Paul Andrew Churnock

Tahsin Erdogan

Joshua Benjamin Levinson

Julien Ridoux

Seyed Arash Sadrieh

Leonid Livak

Jason Eslick

Rajkumar Haridoss

Mark Kamichoff

Yuan Yuan

Shivaraj Tenginakai

Raghavendra Vijayanarasimhan Narasimhan

Faisal M Bhamani

Ryan Parfait

Stephen Holden

Alber Edmon Abdelshahid Moines

Shrawan Chittoor Surender

Shriram Srinivasan

Abhishek Vinayak Naik

Chinchu Merine Joseph

A host-to-ToR switch advertisement protocol can be used at a data center network edge to receive IP advertisements from hosts and convey ToR switch availability status (e.g., in service, intentionally out of service, or unintentionally out of service) back to hosts and vice versa. The host server computers can have the ability to advertise an arbitrary number of IPv4 or IPv6 prefixes to the network through the ToR switches. A single host server computer can leverage multiple instances across multiple ports to the same switch or different switches to implement load balancing (e.g., equal cost multipath (ECMP)) and fault tolerance. The host server computers benefit from an increased availability by receiving network traffic via multiple ToR switches, a single ToR switch with multiple links, or a combination of both.

Network rack of host server computers with increased network availability

Systems and methods are provided for highly accurate synchronization of machine instances in a distributed, hosted computing environment to a reference timekeeper. In addition to a general communication network accessible to machine instances, the distributed environment includes a second network dedicated to carrying time information, such as a pulse-per-second (PPS) signal to isolated timing hardware within host computing devices. The isolated timing hardware can use the PPS signal, along with a reference time, to set a hardware clock. The isolated timing hardware can further provide an interface to machine instances that enables the instances to read the time of the hardware clock. This configuration enables many instances can share access to a single reference timekeeper, thus synchronizing those instances to a much higher accuracy than in traditional network-based time protocols.

Virtualized hardware clocks for providing highly accurate time information in hosted machine instances

An optimization agent optimizes configuration parameters of a shared memory buffer for ports of a network device (e.g., router). The optimization agent located on the network device (or remotely) takes an input state from the network device's switching logic, such as the packets transmitted or dropped per port, and determines port buffer settings to optimize the settings to achieve a goal. The buffer settings can include minimum and maximum buffer limits per port and the goal can be to maximize port utilization or minimize port loss. The optimization agent can then take an action by applying the buffer configurations to the switching logic and repeat the process iteratively until the goal is reached.

Configuring of a shared memory buffer in a network device

A power system for multi-input devices with shared reserve power includes a first automatic transfer switch (ATS) and a second ATS each coupled at respective inputs to a primary power system and a reserve power system. An output of one of the ATSs is coupled to first power input of a multi-input electrical device and an output of the other ATS is coupled to a second power input of the multi-input electrical device. When primary power is available, electrical power is fed to both the inputs of the multi-input electrical device and when primary power is not available reserve power is fed to both inputs of the multi-input electrical device. If a component fails between the primary and/or reserve power system and the multi-input electrical device, a full load of the multi-input electrical device is fed from a non-affected one of the inputs of the multi-input electrical device.

Power system for multi-input devices with shared reserve power

Systems and methods for managing configuration of network devices are provided. In one embodiment, a network device management service can obtain relevant device-specific information, deployment information, and network topology data associated with a network device to be provisioned. The network device management service can enable generating, identifying or obtaining configuration metadata based on relevant network topology data as well as other applicable information. The network device management service can further generate configuration instructions to facilitate configuration of network devices based on configuration metadata.

Network topology assisted device provisioning

Technologies are provided for distributing intra-network communications evenly across routers in a multi-tier network. Routers in a middle tier of the network receive data transmissions from a router in a lower tier of the network and determine that a destination of the data transmissions is another router in the lower tier of the network. Instead of routing the data transmissions to the destination router directly, the routers in the middle tier of the network route the data transmissions to other routers in a higher tier of the network. The routers in the higher tier of the network then route the data transmissions back to routers in the middle tier of the network, distributing the data transmissions evenly across routers in the middle tier that are connected to the destination router in the lower tier.

Bouncing router communications in a multi-tier network

Technologies are provided for creating virtual routing tables in routers in a multi-tier network. One or more routers in the network are configured to block other routers in a lower tier of the network from receiving routing updates from one another. By blocking the routing updates, the routers in the higher tier of the network cause the routers in the lower tier of the network to create different virtual routing tables that do not include other routers the lower tier of the network. In such a configuration, routers in the lower tier of the network can be connected to routers in the higher tier of the network with partially overlapping uplink connections without a loss of bandwidth utilization. Since the routers in the lower tier of the network are unaware of the partially overlapping uplink connections, the routers in the lower tier of the network can distribute data transmissions to one another evenly across all their uplink connections to the routers in the higher tier of the network, instead of using the overlapping uplink connections exclusively.

Virtual routing tables for routers in a multi-tier network

A dynamic configuration system can manage and configure switches or other network devices that come online in a network. When the dynamic configuration system determines that a network device has come online, the dynamic configuration system can identify the network device (e.g., based on its network location, neighbors, fingerprint, identifier, address or the like), select the appropriate configuration data for the network based on the desired network topology, and transmit the configuration data to the network device. The network device can then load the configuration data and function as a component of the desired network topology.

Dynamic network device configuration

Methods, systems, and devices are described for transmitting and storing routing path information and routing topology information using a single protocol. In particular embodiments, routing path information for an exterior gateway protocol ('EGP') network such as a border gateway protocol ('BGP') network is transmitted throughout an interior gateway protocol (&#x201c;IGP&#x201d;) network using IGP messages such as open-shortest-path-first (&#x201c;OSPF&#x201d;) or intermediate state to intermediates state (&#x201c;IS-IS&#x201d;) messages. The IGP networks may transmit and store the BGP information using type length values (TLVs). As a result, network areas running an IGP may maintain BGP information throughout the network area without the overhead of an iBGP mesh and related message-passing.

Transmission and storage of routing path information and routing topology information using a single routing protocol

A plug may be coupled with a cable and include a fiber ferrule extending from the plug. A biasing mechanism may be arranged to bias a cover toward a covered configuration in which the fiber ferrule is situated within an internal volume defined by the cover and away from an uncovered configuration in which the fiber ferrule is situated at least partially outside of the internal volume defined by the cover. Bristles or other blockers may be positioned along or within a boundary of the internal volume of the cover, may be arranged to block particulate entry through the blockers into the internal volume of the cover in the covered configuration, and may be movable to permit passage of the fiber ferrule through the blockers in response to movement of the cover between the covered configuration and uncovered configuration.

Contamination prevention system for fiber optic cabling

Approaches are described for managing configuration information updates, or other information in a data center, shared resource environment, or other such electronic environment. A management component is provided that can manage distribution of configuration information, operating system (OS) images, authentication data, or other such information for host machines, servers, or other such devices in a network environment. The management component can further perform a number of actions when deploying configuration information updates, such as performing one or more checks to verify the health of a target switch, deactivating the target switch, shifting traffic to and from the target switch, updating a current configuration of the target switch, verifying the target switch's health after the configuration changes, bringing the target switch back in service, among others.

Managing configuration updates

Network addresses are dynamically allocated to devices based on demand. It is determined that a quantity of virtualized machine instances executed on a plurality of computing devices has changed. An updated network address allocation is dynamically assigned to a subdivision of a network in response to the quantity of virtualized machine instances. The subdivision includes the plurality of computing devices. A router for the subdivision is reconfigured to implement the updated network address allocation.

Dynamically allocating network addresses

Address resolution in an unnumbered, pseudo-point-to-point network utilizes address transmissions, such as an address advertisement or an address response, in order to obtain address information for use in frame addressing. In one embodiment, routers communicate using a multi-access data link layer protocol, such as Ethernet, but in a physical configuration which restricts data link layer communications to going between only two nodes, thereby negating the multi-access application of the protocol. With only one possible terminal node, address space is conserved by use of unnumbered network interfaces.

Address resolution in unnumbered pseudo-point-to-point network

Growing community of inventors

Dublin, Ireland

Colin John Whittaker