Facebook, Inc.

Jason M Taylor

Narsing Krishna Vijayrao

Hongzhong Jia

Xiaojun Liang

Per W Brashers

Hans-Juergen Schmidtke

Yael Gregory Maguire

Amir Meir Michael

Dhrubajyoti Borthakur

Eitan Frachtenberg

Song Liu

Jon Brian Ehlen

Yan Zhao

Yusuf Abdulghani

Ming Ni

Mark Douglas Callaghan

Domas Mituzas

A routing circuit coupled to one or more ingress ports and one or more egress ports is configured to receive data packets from the one or more ingress ports and route the data packets to the one or more egress ports based on a header portion of the data packets.

Pre-routing device for data centers

A first server device includes a server processor configured to perform a server processing task. The first server device also includes a network switch. The network switch of the first server device is directly connected to a higher level network switch. The network switch of the first server device is directly connected to a second server device that includes another server processor and another network switch.

Distributed network switches of data centers

A method for performing client-side content inference may include (1) receiving a request to upload, from a client-side device to a server-side device, a content item that includes a first sequence of bytes and a second sequence of bytes, (2) identifying a model configured to output a classification for sequences of bytes, (3) using, at the client-side device, the model to derive a first classification for the first sequence, (4) using, at the client-side device, the model to derive a second classification for the second sequence, and (5) uploading, in response to the request, the content item to the server-side device by (a) uploading the first sequence, (b) uploading the first classification substantially contemporaneous with uploading the first sequence, (c) uploading the second sequence, and (d) uploading the second classification substantially contemporaneous with uploading the second sequence. Various other methods, systems, and computer-readable media are also disclosed.

Systems and methods for performing client-side content inference

Data center systems are described. A router can receive a data packet having a header indicating the destination for that data packet. That router can then route the data packet to another, second router. The routers can consider different portions of the header when making their routing decisions. For example, the different portions can be different sizes to allow for the first router to perform less complex routing decisions while the second router can perform more complex routing decisions.

Embodiments are described for dynamically responding to demand for server computing resources. The embodiments can monitor performance of each of multiple computing systems in a data center, identify a particular computing system of the multiple computing systems for allocation of additional computing power, determine availability of an additional power supply to allocate to the identified computing system, determine availability of a capacity on a power distribution line connected to the particular computing system to provide the additional power supply to the particular computing system, and allocate the additional computing power to the identified computing system as a function of the determined availability of the additional power supply and the determined availability of the capacity on the power distribution line.

Dynamically responding to demand for server computing resources

A method of operating a data transport system on a computing device is disclosed. The method comprises: writing outgoing data in a first memory space on a memory module of a computing device; detecting the outgoing data on the first memory space by a data channel component coupled to the memory module, wherein the first memory space is designated for external data transmission; and generating a transmission signal encoding the outgoing data, via the data channel component, for transmission from the memory module through an inter-device interconnect to an external memory module.

Inter-device data-transport via memory channels

Systems and methods for optical communication using single source optical transmission are disclosed. A representative method includes receiving optical signals at a first frequency by a receiver (RX) from an optical fiber. The received optical signals are routed to a modulator that modulates the optical signals at a second frequency. The optical signals are emitted by a transmitter (TX) back to the optical fiber at the second frequency.

Single source optical transmission

Optical communication using optical resonators with noise margins is disclosed. A representative system includes an optical fiber for transmitting optical signals, a receiver configured to receive the optical signals, and a plurality of optical resonators optically connecting the optical fiber to the receiver. The individual optical resonators can have peak sensitivities at mutually different wavelengths of light. In some embodiments the optical resonators can be Q-switches.

Optical resonators having noise margins

A server platform with integrated power supply including a cubby chassis having one or more partitions defining a plurality of sled positions each configured to receive a server sled. A power supply unit is positioned in the cubby chassis and transversely oriented with respect to the one or more partitions. A power supply harness extends between the power supply unit and each of the plurality of sled positions. Therefore, the server platform can be easily connected to a power source without being mounted in a rack, thereby facilitating bench top testing of a server or servers mounted in the cubby chassis.

Server platform with integrated power supply

The disclosure is directed to storing data in different tiers of a database based on the access pattern of the data. Immutable data, e.g., data that does not change or changes less often than a specified threshold, is stored in a first storage tier of the database, and mutable data, e.g., data that changes more often than immutable data, is stored in a second storage tier of the database. The second storage tier of the database is more performant than the first storage tier, e.g., the second storage tier has a higher write endurance and a lower write latency than the first storage tier. All writes to the database are performed at the second storage tier and reads on both storage tiers. The storage tiers are synchronized, e.g., the set of data is copied from the second to the first storage tier based on a trigger, e.g., a specified schedule.

Tiered data storage architecture

A file storage system may be implemented by coordinating non-exhaustible and exhaustible storage devices. The exhaustible storage devices may be lower grade flash based storage devices. The non-exhaustible storage devices may be storage devices based on magnetic recording mechanisms or higher grade flash storage. The file storage system may store received content items on the exhaustible storage devices. The file storage system may additionally store metadata for the content items and/or the file storage system on the non-exhaustible storage devices. During operation, the file storage system may globally optimize the content items on the exhaustible storage devices such that more heavily accessed items are moved to exhaustible storage devices that have experienced relatively few data operations. The file storage system can move less frequently accessed content items to exhaustible storage devices that have experienced a relatively large number of data operations. As such, the operation lifetime of the exhaustible storage devices as a whole can be prolonged.

File storage system based on coordinated exhaustible and non-exhaustible storage

Some embodiments includes an interrupt-driven data transport architecture utilizing a memory channel bus. For example, a first logic component at a first computing device can initiate a data access request involving a second logic component at a second computing device. The first logic component can store request information associated with the data access request in a predefined memory space of a memory module connected via a memory channel bus to the first logic component and the second logic component. The first logic component can then generate a request-ready interrupt signal through one or more redundant pins of the memory channel bus. The second logic component can be triggered by the interrupt signal to read the request information from the predefined memory space. The second logic component can use that information to complete the request.

Interrupt driven memory signaling

Channel performance can be improved in a storage device, such as a flash memory or a flash-based solid state drive, by dynamically provisioning available data channels for both write and read operations. In one aspect, a set of available data channels on a storage device is partitioned into a set of write channels and a set of read channels according to a read-to-write ratio. Next, when an incoming data stream of mixed read requests and write requests arrives at the storage device, the allocated read channels process the read requests on a first group of memory blocks, which does not include garbage collection and write amplification on the first group of memory blocks. In parallel, the allocated write channels process the write requests on a second group of memory blocks, which does include garbage collection and write amplification on the second group of memory blocks.

Allocation of read/write channels for storage devices

Storing data in a cache is disclosed. It is determined that a data record is not stored in a cache. A random value is generated using a threshold value. It is determined whether to store the data record in the cache based at least in part on the generated random value.

Statistical cache promotion

Various embodiments of the present disclosure relate to a cache stickiness index for providing measurable metrics associated with caches of a content delivery networking system. In one embodiment, a method for generating a cache stickiness index, including a cluster stickiness index and a region stickiness index, is disclosed. In embodiments, the cluster stickiness index is generated by comparing cache keys shared among a plurality of front-end clusters. In embodiments, the region stickiness index is generated by comparing cache keys shared among a plurality of data centers. In one embodiment, a system comprising means for generating a stickiness index is disclosed.

Growing community of inventors

Piedmont, CA, United States of America

Jason M Taylor