Showing papers by "Amazon.com published in 2012"

Using virtual networking devices and routing information to initiate external actions

Abstract: Techniques are described for providing managed virtual computer networks that have a configured logical network topology with virtual networking devices, such as by a network-accessible configurable network service, with corresponding networking functionality provided for communications between multiple computing nodes of the virtual computer network by emulating functionality that would be provided by the virtual networking devices if they were physically present. In some situations, the networking functionality provided for a managed computer network of a client includes receiving routing communications directed to the virtual networking devices, and using included routing information to identify and initiate external actions whose effects are not related to how network communications between computing nodes of the managed computer network are configured to be routed or otherwise forwarded through the managed computer network, such as external actions that affect devices that are not part of the managed computer network, or other types of external actions.

Automated profiling of resource usage

Abstract: Operating profiles for consumers of computing resources may be automatically determined based on an analysis of actual resource usage measurements and other operating metrics. Measurements may be taken while a consumer, such as a virtual machine instance, uses computing resources, such as those provided by a host. A profile may be dynamically determined based on those measurements. Profiles may be generalized such that groups of consumers with similar usage profiles are associated with a single profile. Assignment decisions may be made based on the profiles, and computing resources may be reallocated or oversubscribed if the profiles indicate that the consumers are unlikely to fully utilize the resources reserved for them. Oversubscribed resources may be monitored, and consumers may be transferred to different resource providers if contention for resources is too high.

Gaze-based content display

Abstract: A user interface can enable a user to control the display of content in a way that is natural for the user and requires little physical interaction. The gaze direction and/or viewing location of a user can be determined using any of a variety of imaging or other such technologies. By determining the location at which the user is gazing, an electronic device can control aspects such as the scroll rate or page turns of displayed content. In many cases, a device utilizes the natural reading or viewing style of a user to determine appropriate aspects for that user, and can update automatically as conditions change based at least in part upon the change in gaze location and/or viewing patterns.

Managing use of intermediate destination computing nodes for provided computer networks

Abstract: Techniques are described for providing managed computer networks. In some situations, the techniques include managing communications for computing nodes of a managed computer network by using one or more particular computing nodes of the managed computer network that are configured to operate as intermediate destinations to handle at least some communications that are sent by and/or directed to one or more other computing nodes of the managed computer network. For example, a manager module associated with a source computing node may select one or more particular intermediate destination computing nodes to use for one or more particular communications from the source computing node to an indicated final destination, such as based on a configured logical network topology for the managed computer network. The manager module then forwards those communications to a first of the selected intermediate destination computing nodes for further handling.

Optimizing content management

Abstract: A system and method for monitoring the performance associated with fulfilling resource requests and determining optimizations for improving such performance are provided. A processing device obtains and processes performance metric information associated with processing a request for an original resource and any embedded resource. The processing device uses the processed performance metric information to determine a CDN service provider for alternatively hosting at least a portion of the original resource and/or any embedded resources. In some embodiments, in making such a determination, the processing device assesses performance metric information collected and associated with subsequent resource requests for the original resource and any embedded resources using each of a variety of alternative CDN service providers. Aspects of systems and methods for generating recommendations associated with monitoring the operation and performance of CDN service providers with respect to specific resource requests are also provided.

Frameworks and interfaces for offload device-based packet processing

Abstract: High-speed processing of packets to, and from, a virtualization environment can be provided while utilizing hardware-based segmentation offload and other such functionality. A hardware vendor such as a network interface card (NIC) manufacturer can enable the hardware to support open and proprietary stateless tunneling in conjunction with a protocol such as single root I/O virtualization (SR-IOV) in order to implement a virtualized overlay network. The hardware can utilize various rules, for example, that can be used by the NIC to perform certain actions, such as to encapsulate egress packets and decapsulate packets.

Managing content delivery network service providers by a content broker

Abstract: A system, method, and computer readable medium for managing network storage provider and CDN service providers are provided. A content broker component obtains client computing device requests for content provided by a content provider. The content broker processes the client computing device requests and determines whether a subsequent request for the resource should be directed to a network storage provider or a CDN service provider as a function of the updated or processed by the content broker.

Amazon dynamoDB: a seamlessly scalable non-relational database service

Abstract: Reliability and scalability of an application is dependent on how its application state is managed. To run applications at massive scale requires one to operate datastores that can scale to operate seamlessly across thousands of servers and can deal with various failure modes such as server failures, datacenter failures and network partitions. The goal of Amazon DynamoDB is to eliminate this complexity and operational overhead for our customers by offering a seamlessly scalable database service. In this talk, I will talk about how developers can build applications on DynamoDB without having to deal with the complexity of operating a large scale database.

LT codes-based secure and reliable cloud storage service

Abstract: With the increasing adoption of cloud computing for data storage, assuring data service reliability, in terms of data correctness and availability, has been outstanding. While redundancy can be added into the data for reliability, the problem becomes challenging in the “pay-as-you-use” cloud paradigm where we always want to efficiently resolve it for both corruption detection and data repair. Prior distributed storage systems based on erasure codes or network coding techniques have either high decoding computational cost for data users, or too much burden of data repair and being online for data owners. In this paper, we design a secure cloud storage service which addresses the reliability issue with near-optimal overall performance. By allowing a third party to perform the public integrity verification, data owners are significantly released from the onerous work of periodically checking data integrity. To completely free the data owner from the burden of being online after data outsourcing, this paper proposes an exact repair solution so that no metadata needs to be generated on the fly for repaired data. The performance analysis and experimental results show that our designed service has comparable storage and communication cost, but much less computational cost during data retrieval than erasure codes-based storage solutions. It introduces less storage cost, much faster data retrieval, and comparable communication cost comparing to network coding-based distributed storage systems.

Throughput-sensitive redundancy encoding schemes for data storage

Abstract: Techniques for optimizing data storage are disclosed herein. In particular, methods and systems for implementing redundancy encoding schemes with data storage systems are described. The redundancy encoding schemes may be scheduled according to system and data characteristics. The schemes may span multiple tiers or layers of a storage system. The schemes may be generated, for example, in accordance with a transaction rate requirement, a data durability requirement or in the context of the age of the stored data. The schemes may be designed to rectify entropy-related effects upon data storage. The schemes may include one or more erasure codes or erasure coding schemes. Additionally, methods and systems for improving and/or accounting for failure correlation of various components of the storage system, including that of storage devices such as hard disk drives, are described.

Speech recognition power management

Abstract: Power consumption for a computing device may be managed by one or more keywords. For example, if an audio input obtained by the computing device includes a keyword, a network interface module and/or an application processing module of the computing device may be activated. The audio input may then be transmitted via the network interface module to a remote computing device, such as a speech recognition server. Alternately, the computing device may be provided with a speech recognition engine configured to process the audio input for on-device speech recognition.

Early classification on time series

Abstract: In this paper, we formulate the problem of early classification of time series data, which is important in some time-sensitive applications such as health informatics. We introduce a novel concept of MPL (minimum prediction length) and develop ECTS (early classification on time series), an effective 1-nearest neighbor classification method. ECTS makes early predictions and at the same time retains the accuracy comparable with that of a 1NN classifier using the full-length time series. Our empirical study using benchmark time series data sets shows that ECTS works well on the real data sets where 1NN classification is effective.

Managing CDN registration by a storage provider

Abstract: A system, method, and computer readable medium for managing registration, by a network storage provider, of one or more resources with a CDN service provider are provided. A network storage provider storing one or more resources on behalf of a content provider obtains registration information for registering the one or more resources with a CDN service provider. The registration information may include a request to publish one or more resources to a CDN service provider, an identification of the one or more resources, CDN selection criteria provided by the content provider or otherwise selected, and the like. The network storage provider transmits a CDN generation request corresponding to the registration information to the CDN service provider. Then, the network storage provider manages and processes data pursuant to registration of the one or more resources with the CDN service provider.

Request routing utilizing client location information

Abstract: A system, method, and computer-readable medium for request routing based on client location information are provided. A content delivery network service provider receives a DNS query from a client computing device. The DNS query corresponds to a resource identifier for requested content from the client computing device. The content delivery network service provider transmits an alternative resource identifier in response to the client computing device DNS query. The alternative resource identifier is selected as a function of client location information. The client location information is obtained from information corresponding to a mapping of at least partial IP addresses to known locations. The client computing device then issues a second DNS query to the same content delivery network service provider. The content delivery network service provider can then either resolve the second DNS query or transmit another alternative resource identifier.

Optimization of the Receive Filter and Transmit Sequence for Active Sensing

Abstract: This paper discusses the joint design of receive filters and transmit signals for active sensing applications such as radar and active sonar. The goal is to minimize the mean-square error (MSE) of target's scattering coefficient estimate in the presence of clutter and interference, which is equivalent to maximizing the signal-to-clutter-plus-interference ratio. A discrete-time signal model is assumed and practical constant-modulus or low peak-to-average-power ratio (PAR) constraints are imposed on the transmit signal. Several optimization methods are proposed for this joint design. Furthermore, the MSE criterion is expressed in the frequency domain and a corresponding MSE lower bound is derived. Numerical examples for different types of interferences are included to demonstrate the effectiveness of the proposed designs.

Architecture for multi-domain utterance processing

Abstract: Features are disclosed for processing a user utterance with respect to multiple subject matters or domains, and for selecting a likely result from a particular domain with which to respond to the utterance or otherwise take action. A user utterance may be transcribed by an automatic speech recognition (“ASR”) module, and the results may be provided to a multi-domain natural language understanding (“NLU”) engine. The multi-domain NLU engine may process the transcription(s) in multiple individual domains rather than in a single domain. In some cases, the transcription(s) may be processed in multiple individual domains in parallel or substantially simultaneously. In addition, hints may be generated based on previous user interactions and other data. The ASR module, multi-domain NLU engine, and other components of a spoken language processing system may use the hints to more efficiently process input or more accurately generate output.

Plant and animal endemism in the eastern Andean slope: challenges to conservation

Abstract: The Andes-Amazon basin of Peru and Bolivia is one of the most data-poor, biologically rich, and rapidly changing areas of the world. Conservation scientists agree that this area hosts extremely high endemism, perhaps the highest in the world, yet we know little about the geographic distributions of these species and ecosystems within country boundaries. To address this need, we have developed conservation data on endemic biodiversity (~800 species of birds, mammals, amphibians, and plants) and terrestrial ecological systems (~90; groups of vegetation communities resulting from the action of ecological processes, substrates, and/or environmental gradients) with which we conduct a fine scale conservation prioritization across the Amazon watershed of Peru and Bolivia. We modelled the geographic distributions of 435 endemic plants and all 347 endemic vertebrate species, from existing museum and herbaria specimens at a regional conservation practitioner's scale (1:250,000-1:1,000,000), based on the best available tools and geographic data. We mapped ecological systems, endemic species concentrations, and irreplaceable areas with respect to national level protected areas. We found that sizes of endemic species distributions ranged widely ( 200,000 km2) across the study area. Bird and mammal endemic species richness was greatest within a narrow 2500-3000 m elevation band along the length of the Andes Mountains. Endemic amphibian richness was highest at 1000-1500 m elevation and concentrated in the southern half of the study area. Geographical distribution of plant endemism was highly taxon-dependent. Irreplaceable areas, defined as locations with the highest number of species with narrow ranges, overlapped slightly with areas of high endemism, yet generally exhibited unique patterns across the study area by species group. We found that many endemic species and ecological systems are lacking national-level protection; a third of endemic species have distributions completely outside of national protected areas. Protected areas cover only 20% of areas of high endemism and 20% of irreplaceable areas. Almost 40% of the 91 ecological systems are in serious need of protection (= < 2% of their ranges protected). We identify for the first time, areas of high endemic species concentrations and high irreplaceability that have only been roughly indicated in the past at the continental scale. We conclude that new complementary protected areas are needed to safeguard these endemics and ecosystems. An expansion in protected areas will be challenged by geographically isolated micro-endemics, varied endemic patterns among taxa, increasing deforestation, resource extraction, and changes in climate. Relying on pre-existing collections, publically accessible datasets and tools, this working framework is exportable to other regions plagued by incomplete conservation data.

Image analysis for user authentication

Abstract: A user can be authenticated to any of a number of computing devices using an authentication process that recognizes the user and verifies that an actual human being is attempting to be authenticated, in order to minimize the ability of another person to spoof the authentication process. A model of a user can be generated and stored in the cloud, enabling that model to be synchronized across various devices. A user accessing one of these devices can have image information captured, which can be used with a facial recognition process, to recognize the user, and with a human verification process, to verify that the facial information corresponds to a human user. Various approaches such as visual analysis, three-dimensional imaging, and thermal imaging can be used to verify that the human user being recognized is interactive with the device.

Monitoring performance and operation of data exchanges

Abstract: A system and method for monitoring the performance associated with fulfilling resource requests are provided. One or more client computing devices obtain an original resource request and associate a record identifier with the original resource request. The one or more client computing devices also determine performance data associated with processing each embedded resource request included in a response to the original resource request. Each embedded resource request is associated with a component record identifier that is associated with, but different from, the record identifier of the original resource request. The one or more client computing devices can then transmit the determined performance data with the record identifier to another processing device.

Establishing secure remote access to private computer networks

Abstract: Techniques are described for providing users with access to computer networks, such as to enable users to interact with a remote configurable network service to create and configure computer networks that are provided by the configurable network service for use by the users Secure private access between a computer network provided for a user by the configurable network service and one or more other remote computing systems of the user (eg, a remote private network) may be enabled in various ways For example, a user may programmatically invoke an API provided by the configurable network service to obtain assistance in establishing remote access from a remote location to a provided computer network of the configurable network service, such as to establish a VPN connection from the remote location to the provided computer network using hardware and/or software supplied to the remote location in response to the API invocation

Point of presence management in request routing

Abstract: A system and method for the management of client computing device DNS queries and subsequent resource requests within a content delivery network service provider domain are provided. The management of the DNS queries can include the selection of computing devices corresponding to various Point of Presence locations for processing DNS queries. Additionally, the management of the content requests can include the selection of computing devices corresponding to resource cache components corresponding to various Point of Presence locations for providing requested content. The selection of the computing devices can incorporate logic related to geographic criteria, testing criteria, and the like.

Gesture recognition for device input

Abstract: A user can make a symbol with their hand, or other such gesture, at a distance from a computing device that can be captured by at least one imaging element of the device. The captured information can be analyzed to attempt to determine the location of distinguishing features of the symbol in the image information. The image information is then compared to hand gesture information stored in, for example, a library of hand gestures for the user. Upon identifying a match, an input to an application executing on the computing device is provided when the image information contains information matching at least one hand gesture with at least a minimum level of certainty. The hand gesture could include a single “static” gesture, such as a specific letter in sign language, for example, or include two or more “static” gestures. The gesture could also include motion, such as hand movement.

Speech model retrieval in distributed speech recognition systems

Abstract: Features are disclosed for managing the use of speech recognition models and data in automated speech recognition systems. Models and data may be retrieved asynchronously and used as they are received or after an utterance is initially processed with more general or different models. Once received, the models and statistics can be cached. Statistics needed to update models and data may also be retrieved asynchronously so that it may be used to update the models and data as it becomes available. The updated models and data may be immediately used to re-process an utterance, or saved for use in processing subsequently received utterances. User interactions with the automated speech recognition system may be tracked in order to predict when a user is likely to utilize the system. Models and data may be pre-cached based on such predictions.

User identification by gesture recognition

Abstract: A user can be identified and/or authenticated to an electronic device by analyzing aspects of a motion or gesture made by that user. At least one imaging element of the device can capture image information including the motion or gesture, and can determine time-dependent information about that motion or gesture in two or three dimensions of space. The time-dependent information can be used to identify varying speeds, motions, and other such aspects that are indicative of a particular user. The way in which a gesture or motion is made, in addition to the motion or gesture itself, can be used to authenticate an individual user. While other persons can learn the basic gesture or motion, the way in which each person makes that gesture or motion will generally be at least slightly different, which can be used to prevent unauthorized access to sensitive information, protected functionality, or other such content.

Data storage integrity validation

Abstract: Embodiments of the present disclosure are directed to, among other things, validating the integrity of received and/or stored data payloads In some examples, a storage service may perform a first partitioning of a data object into first partitions based at least in part on a first operation The storage service may also verify the data object, by utilizing a verification algorithm, to generate a first verification value In some cases, the storage service may additionally perform a second partitioning of the data object into second partitions based at least in part on a second operation The second partitions may be different from the first partitions Additionally, the archival data storage service may verify the data object using the verification algorithm to generate a second verification value Further, the storage service may determine whether the second verification value equals the first verification value

Filling an order at an inventory pier

Abstract: A system includes an inventory pier, a mobile drive unit, and a management module. The inventory pier includes a defined area arranged to station an inventory holder. The inventory holder stores an inventory item. The mobile drive unit is operable to transport an order holder. The management module is operable to calculate a metric associated with demand for the inventory item. Based at least in part upon the metric, the management module is operable to select, from a plurality of inventory holders, the inventory holder storing the inventory item to be stationed at the defined area of the inventory pier. The management module is further operable to receive an order for the inventory item and instruct the mobile drive unit to transport the order holder to the inventory pier. The order holder is operable to receive the inventory item from the inventory holder proximate to the defined area.

Characterizing environment using ultrasound pilot tones

Abstract: A voice-activated computing device configured to transmit a pilot tone and then capture or receive a signal, which corresponds to the pilot tone, reflected from within the environment containing the voice-activated computing device. The voice-activated computing device, or some other computing system or device, analyzes the received signal in order to determine analyze one or more characteristics present within the signal, i.e. noise, echo, etc. Based upon the analysis, models for signal processing can be determined, selected and/or altered. Future signals received by the voice-activated computing device can be processed with such models. The analysis can also allow for the models to be dynamically updated and for models to be dynamically created.

Modeling the spatial and temporal heterogeneity of deforestation‐driven carbon emissions: the INPE‐EM framework applied to the Brazilian Amazon

Abstract: We present a generic spatially explicit modeling framework to estimate carbon emissions from deforestation (INPE-EM). The framework incorporates the temporal dynamics related to the deforestation process and accounts for the biophysical and socioeconomic heterogeneity of the region under study. We build an emission model for the Brazilian Amazon combining annual maps of new clearings, four maps of biomass, and a set of alternative parameters based on the recent literature. The most important results are as follows: (a) Using different biomass maps leads to large differences in estimates of emission; for the entire region of the Brazilian Amazon in the last decade, emission estimates of primary forest deforestation range from 0.21 to 0.26 Pg C yr � 1 . (b) Secondary vegetation growth presents a small impact on emission balance because of the short duration of secondary vegetation. In average, the balance is only 5% smaller than the primary forest deforestation emissions. (c) Deforestation rates decreased significantly in the Brazilian Amazon in recent years, from 27 Mkm 2 in 2004 to 7 Mkm 2 in 2010. INPE-EM process-based estimates reflect this decrease even though the agricultural frontier is moving to areas of higher biomass. The decrease is slower than a non-process instantaneous model would estimate as it considers residual emissions (slash, wood products, and secondary vegetation). The average balance, considering all biomass, decreases from 0.28 in 2004 to 0.15 Pg C yr � 1 in 2009; the nonprocess model estimates a decrease from 0.33 to 0.10 Pg C yr � 1 . We conclude that the INPE-EM is a powerful tool for representing deforestation-driven carbon emissions. Biomass estimates are still the largest source of uncertainty in the effective use of this type of model for informing mechanisms such as REDD+. The results also indicate that efforts to reduce emissions should focus not only on controlling primary forest deforestation but also on creating incentives for the restoration of secondary forests.