STATEFUL RESOURCE POOL MANAGEMENT FOR JOB EXECUTION

BACKGROUND

Computing systems for querying of large sets of data can be extremely difficult to implement and maintain. In many scenarios, for example, it is necessary to first create and configure the infrastructure (e.g. server computers, storage devices, networking devices, etc.) to be used for the querying operations. It might then be necessary to perform extract, transform, and load (“ETL”) operations to obtain data from a source system and place the data in data storage. It can also be complex and time consuming to install, configure, and maintain the database management system (“DBMS”) that performs the query operations. Moreover, many DBMS are not suitable for querying extremely large data sets in a performant manner.

Computing clusters can be utilized in some scenarios to query large data sets in a performant manner. For instance, a computing cluster can have many nodes that each execute a distributed query framework for performing distributed querying of a large data set. Such computing clusters and distributed query frameworks are, however, also difficult to implement, configure, and maintain. Moreover, incorrect configuration and/or use of computing clusters such as these can result in the non-optimal utilization of processor, storage, network and, potentially, other types of computing resources.

The disclosure made herein is presented with respect to these and other considerations.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates a logical block diagram of stateful resource pool management for job execution, according to some embodiments.

FIG. 3 is a logical block diagram illustrating a managed query service, according to some embodiments.

FIG. 4 is a diagram illustrating interactions between clients and managed query service, according to some embodiments.

FIG. 5 is a sequence diagram for managed execution of queries, according to some embodiments.

FIG. 6 is a sequence diagram for managed execution of queries utilizing a resource planner, according to some embodiments.

FIG. 7 is a logical block diagram illustrating a cluster processing a query as part of managed query execution, according to some embodiments.

FIG. 8 is a logical block diagram illustrating a resource management service, according to some embodiments.

FIG. 9 is logical block diagram illustrating interactions between a resource management service and pools of resources, according to some embodiments.

FIG. 10 is a state diagram illustrating different resource pool states tracked by a resource manager service, according to some embodiments.

FIG. 11 is a state diagram illustrating different computing resource states, according to some embodiments according to some embodiments.

FIG. 12 is a high-level flowchart illustrating various methods and techniques to implement stateful management of resources pools executing jobs, according to some embodiments.

FIG. 13 is a high-level flowchart illustrating techniques to monitor a resource pool for modification events according to some embodiments.

FIGS. 14A-4C describe various techniques for managing a pool of computing resources for executing queries, according to some embodiments.

FIG. 15 is a logical block diagram that shows an illustrative operating environment that includes a service provider network that can be configured to implement aspects of the functionality described herein, according to some embodiments.

FIG. 16 is a logical block diagram illustrating a configuration for a data center that can be utilized to implement aspects of the technologies disclosed herein, according to some embodiments.

FIG. 17 illustrates an example system configured to implement the various methods, techniques, and systems described herein, according to some embodiments.

While embodiments are described herein by way of example for several embodiments and illustrative drawings, those skilled in the art will recognize that embodiments are not limited to the embodiments or drawings described. It should be understood, that the drawings and detailed description thereto are not intended to limit embodiments to the particular form disclosed, but on the contrary, the intention is to cover all modifications, equivalents and alternatives falling within the spirit and scope as defined by the appended claims. The headings used herein are for organizational purposes only and are not meant to be used to limit the scope of the description or the claims. As used throughout this application, the word “may” is used in a permissive sense (i.e., meaning having the potential to), rather than the mandatory sense (i.e., meaning must). Similarly, the words “include,” “including,” and “includes” mean including, but not limited to.

It will also be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first contact could be termed a second contact, and, similarly, a second contact could be termed a first contact, without departing from the scope of the present invention. The first contact and the second contact are both contacts, but they are not the same contact.

DETAILED DESCRIPTION OF EMBODIMENTS

Various embodiments of a stateful resource pool management for job execution are described herein. FIG. 1 illustrates a logical block diagram of stateful resource pool management for job execution, according to some embodiments. Pool(s) 130 of computing resource(s) 140 may be instantiated, configured, and otherwise prepared for executing different types of job(s) 170 on behalf of network-based service(s) 120, in various embodiments. For example, a query management service, such as discussed below with regard to FIGS. 2-8, may utilize computing resource(s) 140 from different pool(s) 130 in order to execute queries with respect to remotely stored data, in some embodiments. Other types of processing jobs, such as Extract Transform Load (ETL), data validation, log analysis, simulation, numerical analysis, text analysis, machine learning, or other statistical analysis, may be managed, performed, or otherwise executed on behalf of different network-based services, in some embodiments. As the configurations, operations, or requirements of computing resources to execute such job(s) 170 may be costly or time consuming to procure, pool management for job execution resources 110 may provide a dynamically managed set of computing resource(s) 140 in respective pool(s) 130 that are pre-configured and available for executing job(s) 170 without requiring network-based service(s) 120 to directly manage the number of computing resource(s) used by the network-based service(s) 120, in various embodiments.

For instance, as illustrated in FIG. 1, pool management for job execution 110 may create 190 pool(s) of computing resources 140, which may be single or multi-node clusters, virtualized servers, instantiated execution platforms, query engines, processing frameworks, or any other set of one or more resource(s) that can execute job(s) 170 selectively routed to computing resource(s) 140, in one embodiment. Computing resource(s) 140 may interact with other services, data stores, or computing resources (not illustrated), such as accessing remotely stored data, or invoking functions, operations, or processes executed by a separate system, in some embodiments. Pool management for job execution resources 110 may then provide the pools of resources 130 to network-based service(s) 120 for job execution. For example, pool management for job execution 110 may implement an interface, such as discussed below with regard to FIG. 9, via which network-based service(s) 120 can programmatically get resource(s) 150 for executing a job 170, in one embodiment. Pool management for job execution resources 110 may identify a pool 130 and computing resource(s) 140 within the pool to execute the job 170 for the network-based service 120 and provide the resource(s) 160 in response to the request, in one embodiment. For example, pool management for job execution resource(s) may identify a pool 130 specially provisioned for the network-based 120 service or a pool 130 provisioned for the type of job to be executed by the network-based service 120, in one embodiment. Pool management for job execution resources 110 may then randomly assign a resource from the pool, or may deterministically select a resource (e.g., based on characteristics of the computing resource, network-based service, or job), in one embodiment, such as a type of computing resource that implements a particular type of query engine for processing a job that is a query. Once the resource(s) 160 are provided to network-based service(s) 120 (e.g., by providing an identifier or access credential for reaching the resource).

As noted above, pool management for job execution resource(s) 110 may dynamically manage computing resource(s) 140 and pool(s) 130. For example, pool management for job execution resource 110 may collect metric(s) 180 for a pool 130 of computing resources 140. For instance, various kinds of data events for individual computing resource(s) 140, like performance utilization metrics for processor capacity, network-bandwidth, storage capacity, I/O bandwidth, health metrics for the computing resource(s) itself (e.g., start up time) or the environment of the computing resource(s) (e.g., network events), job execution status or state indications, or other information may be provided as part of metric(s) 180, in some embodiments. Pool management for job execution resource(s) 110 may collect, aggregate, and analyze metric(s) 180 for different pools 130, in one embodiment. For example, pool management for job execution resource(s) 110 may determine the average time it takes for a computing resource to clean up, scrub, or otherwise prepare to accept a new job for execution, in one embodiment. Based on such an average time, pool management for job execution resource(s) may increase or decrease the rate at which computing resource(s) 140 are added to pool(s) 130. As discussed below with regard to FIGS. 10-11, pool management for job execution resource(s) 110 may track the state of the pool (e.g., based on event data or other metrics) and the state of resources in the pool (e.g., based on event data or other metric(s), in one embodiment.

Pool management for job execution resources 110 may modify pool(s) 190 based on modification events detected for a pool based on metric(s) 180. For example, based on health or other liveness metrics for computing resource(s) 140 of a given pool, pool management for job execution resource(s) 110 may add or remove computing resource(s) 140 from a pool 130 so that the pool 130 maintains an efficient number of computing resource(s) based on job execution demand for that pool. Modification events may trigger different changes to or affect the operation of a pool 130. For example, modification events may be triggered or detected based on life cycle events or states of a pool 130 (e.g., a pool in warm up state triggers the addition of computing resources, while a pool in a decommission state, triggers the scrubbing or releasing of resources from the pool to discontinue the pool). In some embodiments, modification events may be detected or triggered based on or more modification event criteria (e.g., aggregate metric values for a pool 130 compared to a threshold, a state of the pool or aggregation of state for individual resources in the pool compared to state conditions). Modification event criteria may be determined based on machine learning or other statistical analyses of historical pool metrics or may be specified according to an interface that allows a user of a pool (e.g., an administrator or develop of a network-based service 120) to specify the event criteria for different types of modification events (e.g., values for thresholds that trigger the adding or removing of resources from a pool), in some embodiments.

Please note that the previous description of stateful resource pool management for executing jobs is a logical illustration and thus is not to be construed as limiting as to the implementation of a network-based service, pool of computing resources, pool of computing resources, or pool management for job execution resources.

This specification begins with a general description of a provider network that implements a resource management service that provides stateful pool management for the execution of jobs that are queries received from another network-based service, a managed query service. Then various examples of the managed query service and resource management service (along with other services that may be utilized or implemented) including different components/modules, or arrangements of components/module that may be employed as part of implementing the services are discussed. A number of different methods and techniques to implement stateful pool management for the execution of jobs are then discussed, some of which are illustrated in accompanying flowcharts. Finally, a description of an example computing system upon which the various components, modules, systems, devices, and/or nodes may be implemented is provided. Various examples are provided throughout the specification.

FIG. 2 is a logical block diagram illustrating a provider network offering a resource management service for performing stateful pool management for jobs executed on behalf of other network-based services in the provider network, according to some embodiments. Provider network 200 may be a private or closed system or may be set up by an entity such as a company or a public sector organization to provide one or more services (such as various types of cloud-based storage) accessible via the Internet and/or other networks to clients 250, in some embodiments. Provider network 200 may be implemented in a single location or may include numerous data centers hosting various resource pools, such as collections of physical and/or virtualized computer servers, storage devices, networking equipment and the like (e.g., FIGS. 15, 16 and computing system 2000 described below with regard to FIG. 17), needed to implement and distribute the infrastructure and storage services offered by the provider network 200. In some embodiments, provider network 200 may implement various computing resources or services, such as a virtual compute service 210, data processing service(s) 220, (e.g., relational or non-relational (NoSQL) database query engines, map reduce processing, data flow processing, and/or other large scale data processing techniques), data storage service(s) 230, (e.g., an object storage service, block-based storage service, or data storage service that may store different types of data for centralized access) other services 240 (any other type of network based services (which may include various other types of storage, processing, analysis, communication, event handling, visualization, and security services not illustrated), managed query service 270, data catalog service 280, and resource management service 290.

In various embodiments, the components illustrated in FIG. 2 may be implemented directly within computer hardware, as instructions directly or indirectly executable by computer hardware (e.g., a microprocessor or computer system), or using a combination of these techniques. For example, the components of FIG. 2 may be implemented by a system that includes a number of computing nodes (or simply, nodes), each of which may be similar to the computer system embodiment illustrated in FIG. 17 and described below. In various embodiments, the functionality of a given system or service component (e.g., a component of data storage service 230) may be implemented by a particular node or may be distributed across several nodes. In some embodiments, a given node may implement the functionality of more than one service system component (e.g., more than one data store component).

Virtual compute service 210 may be implemented by provider network 200, in some embodiments. Virtual computing service 210 may offer instances and according to various configurations for client(s) 250 operation. A virtual compute instance may, for example, comprise one or more servers with a specified computational capacity (which may be specified by indicating the type and number of CPUs, the main memory size, and so on) and a specified software stack (e.g., a particular version of an operating system, which may in turn run on top of a hypervisor). A number of different types of computing devices may be used singly or in combination to implement the compute instances and of provider network 200 in different embodiments, including general purpose or special purpose computer servers, storage devices, network devices and the like. In some embodiments instance client(s) 250 or other any other user may be configured (and/or authorized) to direct network traffic to a compute instance.

Compute instances may operate or implement a variety of different platforms, such as application server instances, Java™ virtual machines (JVMs), general purpose or special-purpose operating systems, platforms that support various interpreted or compiled programming languages such as Ruby, Perl, Python, C, C++ and the like, or high-performance computing platforms) suitable for performing client(s) 202 applications, without for example requiring the client(s) 250 to access an instance. Applications (or other software operated/implemented by a compute instance and may be specified by client(s), such as custom and/or off-the-shelf software.

In some embodiments, compute instances have different types or configurations based on expected uptime ratios. The uptime ratio of a particular compute instance may be defined as the ratio of the amount of time the instance is activated, to the total amount of time for which the instance is reserved. Uptime ratios may also be referred to as utilizations in some implementations. If a client expects to use a compute instance for a relatively small fraction of the time for which the instance is reserved (e.g., 30%-35% of a year-long reservation), the client may decide to reserve the instance as a Low Uptime Ratio instance, and pay a discounted hourly usage fee in accordance with the associated pricing policy. If the client expects to have a steady-state workload that requires an instance to be up most of the time, the client may reserve a High Uptime Ratio instance and potentially pay an even lower hourly usage fee, although in some embodiments the hourly fee may be charged for the entire duration of the reservation, regardless of the actual number of hours of use, in accordance with pricing policy. An option for Medium Uptime Ratio instances, with a corresponding pricing policy, may be supported in some embodiments as well, where the upfront costs and the per-hour costs fall between the corresponding High Uptime Ratio and Low Uptime Ratio costs.

Compute instance configurations may also include compute instances with a general or specific purpose, such as computational workloads for compute intensive applications (e.g., high-traffic web applications, ad serving, batch processing, video encoding, distributed analytics, high-energy physics, genome analysis, and computational fluid dynamics), graphics intensive workloads (e.g., game streaming, 3D application streaming, server-side graphics workloads, rendering, financial modeling, and engineering design), memory intensive workloads (e.g., high performance databases, distributed memory caches, in-memory analytics, genome assembly and analysis), and storage optimized workloads (e.g., data warehousing and cluster file systems). Size of compute instances, such as a particular number of virtual CPU cores, memory, cache, storage, as well as any other performance characteristic. Configurations of compute instances may also include their location, in a particular data center, availability zone, geographic, location, etc. . . . and (in the case of reserved compute instances) reservation term length. Different configurations of compute instances, as discussed below with regard to FIG. 3, may be implemented as computing resources associated in different pools of resources managed by resource management service 290 for executing jobs routed to the resources, such as queries routed to select resources by managed query service 270.

Data processing services 220 may be various types of data processing services to perform different functions (e.g., query or other processing engines to perform functions such as anomaly detection, machine learning, data lookup, or any other type of data processing operation). For example, in at least some embodiments, data processing services 230 may include a map reduce service that creates clusters of processing nodes that implement map reduce functionality over data stored in one of data storage services 240. Various other distributed processing architectures and techniques may be implemented by data processing services 230 (e.g., grid computing, sharding, distributed hashing, etc.). Note that in some embodiments, data processing operations may be implemented as part of data storage service(s) 230 (e.g., query engines processing requests for specified data). Data processing service(s) 230 may be clients of data catalog service 220 in order to obtain structural information for performing various processing operations with respect to data sets stored in data storage service(s) 230, as provisioned resources in a pool for managed query service 270.

Data catalog service 280 may provide a catalog service that ingests, locates, and identifies data and the schema of data stored on behalf of clients in provider network 200 in data storage services 230. For example, a data set stored in a non-relational format may be identified along with a container or group in an object-based data store that stores the data set along with other data objects on behalf of a same customer or client of provider network 200. In at least some embodiments, data catalog service 280 may direct the transformation of data ingested in one data format into another data format. For example, data may be ingested into data storage service 230 as single file or semi-structured set of data (e.g., JavaScript Object Notation (JSON)). Data catalog service 280 may identify the data format, structure, or any other schema information of the single file or semi-structured set of data. In at least some embodiments, the data stored in another data format may be converted to a different data format as part of a background operation (e.g., to discover the data type, column types, names, delimiters of fields, and/or any other information to construct the table of semi-structured data in order to create a structured version of the data set). Data catalog service 280 may then make the schema information for data available to other services, computing devices, or resources, such as computing resources or clusters configured to process queries with respect to the data, as discussed below with regard to FIGS. 3-7.

Data storage service(s) 230 may implement different types of data stores for storing, accessing, and managing data on behalf of clients 250 as a network-based service that enables clients 250 to operate a data storage system in a cloud or network computing environment. For example, data storage service(s) 230 may include various types of database storage services (both relational and non-relational) for storing, querying, and updating data. Such services may be enterprise-class database systems that are highly scalable and extensible. Queries may be directed to a database in data storage service(s) 230 that is distributed across multiple physical resources, and the database system may be scaled up or down on an as needed basis. The database system may work effectively with database schemas of various types and/or organizations, in different embodiments. In some embodiments, clients/subscribers may submit queries in a number of ways, e.g., interactively via an SQL interface to the database system. In other embodiments, external applications and programs may submit queries using Open Database Connectivity (ODBC) and/or Java Database Connectivity (JDBC) driver interfaces to the database system.

One data storage service 230 may be implemented as a centralized data store so that other data storage services may access data stored in the centralized data store for processing and or storing within the other data storage services, in some embodiments. A may provide storage and access to various kinds of object or file data stores for putting, updating, and getting various types, sizes, or collections of data objects or files. Such data storage service(s) 230 may be accessed via programmatic interfaces (e.g., APIs) or graphical user interfaces. A centralized data store may provide virtual block-based storage for maintaining data as part of data volumes that can be mounted or accessed similar to local block-based storage devices (e.g., hard disk drives, solid state drives, etc.) and may be accessed utilizing block-based data storage protocols or interfaces, such as internet small computer interface (iSCSI).

In at least some embodiments, one of data storage service(s) 230 may be a data warehouse service that utilizes a centralized data store implemented as part of another data storage service 230. A data warehouse service as may offer clients a variety of different data management services, according to their various needs. In some cases, clients may wish to store and maintain large of amounts data, such as sales records marketing, management reporting, business process management, budget forecasting, financial reporting, website analytics, or many other types or kinds of data. A client's use for the data may also affect the configuration of the data management system used to store the data. For instance, for certain types of data analysis and other operations, such as those that aggregate large sets of data from small numbers of columns within each row, a columnar database table may provide more efficient performance. In other words, column information from database tables may be stored into data blocks on disk, rather than storing entire rows of columns in each data block (as in traditional database schemes).

Managed query service 270, as discussed below in more detail with regard to FIGS. 3-7, may manage the execution of queries on behalf of clients so that clients may perform queries over data stored in one or multiple locations (e.g., in different data storage services, such as an object store and a database service) without configuring the resources to execute the queries, in various embodiments. Resource management service 290, as discussed in more detail below with regard to FIGS. 8-14, may manage and provide pools of computing resources for different services like managed query service 270 in order to execute jobs on behalf the different services, as discussed above with regard to FIG. 1.

Generally speaking, clients 250 may encompass any type of client configurable to submit network-based requests to provider network 200 via network 260, including requests for storage services (e.g., a request to create, read, write, obtain, or modify data in data storage service(s) 240, etc.) or managed query service 270 (e.g., a request to query data in a data set stored in data storage service(s) 230). For example, a given client 250 may include a suitable version of a web browser, or may include a plug-in module or other type of code module that may execute as an extension to or within an execution environment provided by a web browser. Alternatively, a client 250 may encompass an application such as a database application (or user interface thereof), a media application, an office application or any other application that may make use of storage resources in data storage service(s) 240 to store and/or access the data to implement various applications. In some embodiments, such an application may include sufficient protocol support (e.g., for a suitable version of Hypertext Transfer Protocol (HTTP)) for generating and processing network-based services requests without necessarily implementing full browser support for all types of network-based data. That is, client 250 may be an application may interact directly with provider network 200. In some embodiments, client 250 may generate network-based services requests according to a Representational State Transfer (REST)-style network-based services architecture, a document- or message-based network-based services architecture, or another suitable network-based services architecture.

In some embodiments, a client 250 may provide access to provider network 200 to other applications in a manner that is transparent to those applications. For example, client 250 may integrate with an operating system or file system to provide storage on one of data storage service(s) 240 (e.g., a block-based storage service). However, the operating system or file system may present a different storage interface to applications, such as a conventional file system hierarchy of files, directories and/or folders. In such an embodiment, applications may not need to be modified to make use of the storage system service model. Instead, the details of interfacing to the data storage service(s) 240 may be coordinated by client 250 and the operating system or file system on behalf of applications executing within the operating system environment.

Clients 250 may convey network-based services requests (e.g., access requests directed to data in data storage service(s) 240, operations, tasks, or jobs, being performed as part of data processing service(s) 230, or to interact with data catalog service 220) to and receive responses from provider network 200 via network 260. In various embodiments, network 260 may encompass any suitable combination of networking hardware and protocols necessary to establish network-based-based communications between clients 250 and provider network 200. For example, network 260 may generally encompass the various telecommunications networks and service providers that collectively implement the Internet. Network 260 may also include private networks such as local area networks (LANs) or wide area networks (WANs) as well as public or private wireless networks. For example, both a given client 250 and provider network 200 may be respectively provisioned within enterprises having their own internal networks. In such an embodiment, network 260 may include the hardware (e.g., modems, routers, switches, load balancers, proxy servers, etc.) and software (e.g., protocol stacks, accounting software, firewall/security software, etc.) necessary to establish a networking link between given client 250 and the Internet as well as between the Internet and provider network 200. It is noted that in some embodiments, clients 250 may communicate with provider network 200 using a private network rather than the public Internet.

FIG. 3 is a logical block diagram illustrating a managed query service, according to some embodiments. As discussed below with regard to FIGS. 4-9, managed query service 270 may leverage the capabilities of various other services in provider network 200. For example, managed query service 270 may utilize resource management service 290 to provision and manage pools of preconfigured resources to execute queries, provide resources of preconfigured queries, and return utilized resources to availability. For example, resource management service 290 may instantiate, configure, and provide resource pool(s) 350a and 350n that include pool resource(s) 352a and 352n from one or more different resource services, such as computing resource(s) 354 in virtual compute service 210 and computing resource(s) 356 in data processing service(s) 220. Resource management service 290 may send requests to create, configure, tag (or otherwise associate) resources 352 for a particular resource pool, terminate, reboot, otherwise operate resources 352 in order to execute jobs on behalf of other network-based services.

Once a resource from a pool is provided (e.g., by receiving an identifier or other indicator of the resource to utilize), managed query service 270 may interact directly with the resource 354 in virtual compute service 210 or the resource 356 in data processing services 220 to execute queries, in various embodiments. Managed query service 270 may utilize data catalog service 280, in some embodiments to store data set schemas 352, as discussed below with regard to FIGS. 4, for subsequent use when processing queries, as discussed below with regard to FIGS. 5-7, in some embodiments. For example, a data set schema may identify the field or column data types of a table as part of a table definition so that a query engine (executing on a computing resource), may be able to understand the data being queried, in some embodiments. Managed query service 270 may also interact with data storage service(s) 230 to directly source data sets 370 or retrieve query results 380, in some embodiments.

Managed query service 270 may implement a managed query interface 310 to handle requests from different client interfaces, as discussed below with regard to FIG. 4. For example, different types of requests, such as requests formatted according to an Application Programmer Interface (API), standard query protocol or connection, or requests received via a hosted graphical user interface implemented as part of managed query service may be handled by managed query interface 310.

Managed query service 270 may implement managed query service control plane 320 to manage the operation of service resources (e.g., request dispatchers for managed query interface 310, resource planner workers for resource planner 330, or query tracker monitors for query tracker 340). Managed query service control plane 320 may direct requests to appropriate components as discussed below with regard to FIGS. 5 and 6. Managed query service 270 may implement authentication and authorization controls for handling requests received via managed query interface 310. For example, managed query service control plane 320 may validate the identity or authority of a client to access the data set identified in a query received from a client (e.g., by validating an access credential). In at least some embodiments, managed query service control plane 320 may maintain (in an internal data store or as part of a data set in an external data store, such as in one of data storage service(s) 230), query history, favorite queries, or query execution logs, and other managed query service historical data. Query execution costs may be billed, calculated or reported by managed query service control plane 320 to a billing service (not illustrated) or other system for reporting usage to users of managed query service, in some embodiments.

Managed query service 270 may implement resource planner 330 to intelligently select available computing resources from pools for execution of queries, in some embodiments. For example, resource planner 330 may evaluated collected data statistics associated with query execution (e.g., reported by computing resources) and determine an estimated number or configuration of computing resources for executing a query within some set of parameters (e.g., cost, time, etc.). For example, machine learning techniques may be applied by resource planner 330 to generate a query estimation model that can be applied to the features of a received query to determine the number/configuration of resources, in one embodiment. Resource planner 330 may then provide or identify which ones of the resources available to execute the query from a pool may best fit the estimated number/configuration, in one embodiment.

In various embodiments, managed query service 270 may implement query tracker 340 in order to manage the execution of queries at compute clusters, track the status of queries, and obtain the resources for the execution of queries from resource management service 290. For example, query tracker 340 may maintain a database or other set of tracking information based on updates received from different managed query service agents implemented on provisioned computing resources (e.g., computing clusters as discussed below with regard to FIGS. 5-7). In some embodiments, query tracker may

FIG. 4 is a diagram illustrating interactions between clients and managed query service, according to some embodiments. Client(s) 400 may be client(s) 250 in FIG. 2 above or other clients (e.g., other services systems or components implemented as part of provider network 200 or as part of an external service, system, or component, such as data exploration or visualization tools (e.g., Tableau, Looker, MicroStrategy, Qliktech, or Spotfire). Clients 400 can send various requests to managed query service 270 via managed query interface 310. Managed query interface 310 may offer a management console 440, which may provider a user interface to submit queries 442 (e.g., graphical or command line user interfaces) or register data schemas 444 for executing queries. For example, management console 440 may be implemented as part of a network-based site (e.g., an Internet website for provider network 200) that provides various graphical user interface elements (e.g., text editing windows, drop-down menus, buttons, wizards or workflows) to submit queries or register data schemas. Managed query interface 310 may implement programmatic interfaces 410 (e.g., various Application Programming

Interface (API) commands) to perform queries, and various other illustrated requests. In some embodiments, managed query interface 310 may implement custom drivers that support standard communication protocols for querying data, such as JDBC driver 430 or ODBC driver 420.

Clients 400 can submit many different types of request to managed query interface 310. For example, in one embodiment, clients 400 can submit requests 450 to create, read, modify, or delete data schemas. For example, a new table schema can be submitted via a request 450. Request 450 may include a name of the data set (e.g., table), a location of the data set (e.g. an object identifier in an object storage service, such as data storage service 230, file path, uniform resource locator, or other location indicator), number of columns, column names, data types for fields or columns (e.g., string, integer, Boolean, timestamp, array, map, custom data types, or compound data types), data format (e.g., formats including, but not limited to, JSON, CSV, AVRO, ORC, PARQUET, tab delimited, comma separated, as well as custom or standard serializers/desrializers), partitions of a data set (e.g., according to time, geographic location, or other dimensions), or any other schema information for process queries with respect to data sets, in various embodiments. In at least some embodiments, request to create/read/modify/delete data set schemas may be performed using a data definition language (DDL), such as Hive Query Language (HQL). Managed query interface 310 may perform respective API calls or other requests 452 with respect to data catalog service 280, to store the schema for the data set (e.g., as part of table schemas 402). Table schemas 402 may be stored in different formats (e.g., Apache Hive). Note, in other embodiments, managed query service 270 may implement its own metadata store.

Clients 400 may also send queries 460 and query status 470 requests to managed query interface 310 which may direct those requests 460 and 470 to managed query service control plane 320, in various embodiments, as discussed below with regard to FIGS. 5 and 6. Queries 460 may be formatted according to various types of query languages, such as Structured Query Language (SQL) or HQL.

Client(s) 400 may also submit requests for query history 480 or other account related query information (e.g., favorite or common queries) which managed query. In some embodiments, client(s) 400 may programmatically trigger the performance of past queries by sending a request to execute a saved query 490, which managed query service control plane 320 may look-up and execute. For example, execute saved query request may include a pointer or other identifier to a query stored or saved for a particular user account or client. Managed query service control plane 320 may then access that user query store to retrieve and execute the query (according to techniques discussed below with regard to FIGS. 5-7).

FIG. 5 is a sequence diagram for managed execution of queries, according to some embodiments. Query 530 may be received at managed query service control plane 320 which may submit the query 532 to query tracker 340 indicating the selected cluster 536 for execution. Query tracker 340 may lease a cluster 534 from resource management service 290, which may return a cluster 536. Resource management service 290 and query tracker 340 may maintain lease state information for resources that are leased by query tracker and assigned to execute received queries. Query tracker 340 may then initiate execution of the query 538 at the provisioned cluster 510, sending a query execution instruction to a managed query agent 512.

Managed query agent 512 may get schema 540 for the data sets(s) 520 from data catalog service 280, which may return the appropriate schema 542 (e.g., implementing a query processing technique that applies schema on-read of data from data set(s)). Provisioned cluster 510 can then generate a query execution plan and execute the query 544 with respect to data set(s) 520 according to the query plan. Managed query agent 512 may send query status 546 to query tracker 340 which may report query status 548 in response to get query status 546 request, sending a response 550 indicating the query status 550. Provisioned cluster 510 may store the query results 552 in a result store 522 (which may be a data storage service 230). Managed query service control plane 320 may receive q request to get a query results 554 and get query results 556 from results store 522 and provide the query results 558 in response, in some embodiments.

FIG. 6 is a sequence diagram for managed execution of queries utilizing a resource planner, according to some embodiments. Query 630 may be received at managed query service control plane 320 which may submit the query 632 to resource planner 340. Resource planner 340 may analyze the query to determine the optimal cluster to process the query based on historical data for processing queries and available cluster(s) 634 received from resource management service 290. Resource planner 340 may then select a query and submit the query to query tracker 340 indicating the selected cluster 636 for execution. Query tracker 340 may then initiate execution of the query 638 at the provisioned cluster 610, sending a query execution instruction to a managed query agent 612.

Managed query agent 612 may get schema 640 for the data sets(s) 620 from data catalog service 280, which may return the appropriate schema 642. Provisioned cluster 610 can then generate a query execution plan and execute the query 644 with respect to data set(s) 620 according to the query plan. Managed query agent 612 may send query status 646 to query tracker 340 which may report query status 648 in response to get query status 646 request, sending a response 650 indicating the query status 650. Provisioned cluster 610 may store the query results 652 in a result store 622 (which may be a data storage service 230). Managed query service control plane 320 may receive q request to get a query results 654 and get query results 656 from results store 622 and provide the query results 658 in response, in some embodiments.

Different types of computing resources may be provisioned and configured in resource pools, in some embodiments. Single-node clusters or multi-node compute clusters may be one example of a type of computing resource provisioned and configured in resource pools by resource management service 290 to service queries for managed query service 270. FIG. 7 is a logical block diagram illustrating a cluster processing a query as part of managed query execution, according to some embodiments. Cluster 710 may implement a computing node 720 that is a leader node (according to the query engine 724 implemented by cluster 710). In some embodiments, no single node may be a leader node, or the leader node may rotate from processing one query to the next. Managed query agent 722 may be implemented as part of leader node 720 in order to provide an interface between the provisioned resource, cluster 710, and other components of managed query service 270 and resource management service 290. For example, managed query agent 722 may provide further data to managed query service 270, such as the status 708 of the query (e.g. executing, performing I/O, performing aggregation, etc.,) and metrics 706 (e.g., health metrics, resource utilization metrics, cost metrics, length of time, etc.). In some embodiments, managed query agent 722 may provide cluster/query status 708 and execution metric(s) 706 to resource management service 290 (in order to make pool management decisions, such as modification events, lease requests, etc.). For example, managed query agent 722 may indicate cluster status 708 to resource management service 290 indicating that a query has completed and that the cluster 710 is ready for reassignment (or other resource lifecycle operations, as discussed below with regard to FIG. 10).

Leader node 720 may implement query engine 724 to execute queries, such as query 702 which may be received via managed query agent 722 as query 703. For instance, managed query agent may implement a programmatic interface for query tracker to submit queries (as discussed above in FIGS. 5 and 6), and then generate and send the appropriate query execution instruction to query engine 724. Query engine 724 may generate a query execution plan for received queries 703. In at least some embodiments, leader node 720, may obtain schema information for the data set(s) 770 from the data catalog service 280 or metadata stores for data 762 (e.g., data dictionaries, other metadata stores, other data processing services, such as database systems, that maintain schema information) for data 762, in order to incorporate the schema data into the generation of the query plan and the execution of the query. Leader node 720 may generate and send query execution instructions 740 to computing nodes that access and apply the query to data 762 in data store(s) 760. Compute nodes, such as nodes 730a, 730b, and 730n, may respectively implement query engines 732a, 732b, and 732n to execute the query instructions, apply the query to the data 750, and return partial results 740 to leader node 720, which in turn may generate and send query results 704. Query engine 724 and query engines 732 may implement various kinds of distributed query or data processing frameworks, such as the open source Presto distributed query framework or the Apache Spark framework.

FIG. 8 is a logical block diagram illustrating a resource management service, according to some embodiments. Resource management service 290 may be responsible for managing resource pools, such as pools of resources that can execute queries on behalf of managed query service 270, or pools of resources configured to execute other types of jobs for other services (e.g., other services in provider network 200). For example, resource management service 290 may implement a lease framework that assigns, allocates, or otherwise leases a computing resource, such as cluster 710 in FIG. 7, to a requesting network-based service (e.g., when query tracker 340 submits a request for a lease to a cluster to execute a query). Resource management service 290 may track the leases of the different resources in the pools as part of resource state 860. Resource management service 290 may instantiate, configure and/or provision resource pools and monitor their health.

Resource management service 290 may track the state of a pool 850 in order to detect modification events for a pool. For example, a resource pool may have a maximum pool size, a minimum number of idle resources in the pool, and a maximum number of idle resources in the pool, which may trigger modification events, as discussed below with regard to FIG. 10. Resource management service 290 may monitor the state of resources 860 according to metrics and other information collected from managed query agents. For example, managed query agent 722 on clusters 710 in FIG. 7 may detect life cycle or other pool management events and send them to the resource management service 290. Once a leased resource is returned, resource management service 290 may check or ensure that the returned resource does not have any state related to previous job executions and can be safely used for another job. Resource management service 290 may direct reuse operations, such as a scrubbing operation, as discussed below with regard to FIG. 14C.

Resource management service 290 may implement metric collection 810, in various embodiments, to handle metrics and other data events received for computing resources in pools. For example, cluster and query execution state reported from clusters can be processed by a group of metrics collection workers that categorize and store metrics according to resource pool. In some embodiments, metric collection 810 may perform initial metrics processing, generating aggregated data values for a pool and storing them for evaluation by pool monitoring 820. In some embodiments, performance metric collection may poll or ping resources (e.g., by sending a status request to a managed query agent) to check for liveness if no metrics have been recorded for the resource within some period of time. Performance metrics for resources may be recorded in metrics store 870 (which may be internally implemented or externally implemented data store (e.g., in storage service 230). State information for resources, including reported state/lifecycle changes discussed below with regard to FIG. 10, may be stored in resource state 860.

Resource management service 290 may monitor resource pools for changes, as discussed below with regard to FIGS. 11-14C. For example, one or more modification events may be defined for a resource pool based on one or more modification event criteria, in some embodiments. Modification event criteria may include numbers of resources in a particular state (e.g., pending, ready, scrubbing, terminated, failed, etc.) compared to thresholds, whether a resource is in a particular state, whether a pool is in a particular state (e.g., warm up, available, cool down, decommission). Pool monitoring 820 may, in some embodiments, generate pool-focused statistics, such as average operation times, median operation times, etc., which may be evaluated as part of event criteria (e.g., average pool startup time exceed X threshold). In at least some embodiments, pool monitoring 820 may access resource state 860 or performance metric(s) 870 to detect a modification event (e.g., checking whether network bandwidth utilization has increased across a group of resources in a pool).

Resource management service 290 may implement pool management 840 to create and decommission pools, instantiate, tag, and configure resources within pools, and remove, migrate, or otherwise modify resources within pools, in various embodiments. For example, pool management 840 may generate and send various requests to other network-based services (e.g., virtual compute service 210 and data processing service 220) to launch, configure, or otherwise provision computing resources, in one embodiment. The launched computing resources may be treated and managed as a pool by resource management service 270 so that virtual compute service 210 and data processing service 220 are unaware of which resource belongs to which pool. Pool management may access resource configuration(s) 880 to determine how resources should be configured. For example, a machine image, resource launch template, configuration script, or other information to configure computing resources for a pool may be maintained in resource configuration(s) 880, which pool management 840 may use to instantiate and configure the computing resources (e.g., in the other services), in one embodiment.

Pool management 840 may perform modification events to resource pools based on the modification events detected by pool monitoring 820. For example, modification events to add or remove computing resources in a resource pool, reconfigure existing resources, migrate resources from one pool to another or any other modification event, as discussed below with regard to FIGS. 11-14C may be performed by pool management 840 performing corresponding requests or operations at other services (e.g., API requests to launch more resources, terminate existing resources, pause resources, reconfigure resources, etc.).

Resource management service 290 may implement resource assignment 830 to handle requests for resources from resource pools, in some embodiments. For example, resource assignment 830 may identify a pool for a resource request (e.g., based on an identifier included in the request, such as a pool identifier or user identifier). Resource assignment 830 may then select a resource from the pool (e.g., randomly selecting an available resource) or according to a deterministic selection technique (e.g., a queue of resources). In at least some embodiments, resource assignment 830 may utilize a leasing scheme, granting resources to leases in a pool for a period of time. For example, resource assignment may track lease states for resources. Lease states may include ongoing (e.g., a job is executing), expired (e.g., the time is up and the resource may be forcibly returned), or terminated (e.g., job completed prior to lease expiration), in some embodiments. Some modification events for a cluster, such as returning a cluster to the pool, may be triggered based on a change in lease state (e.g., a change to expired or terminated). Lease states may be maintained as part of resource state store 860, in some embodiments.

FIG. 9 is logical block diagram illustrating interactions between a resource management service and pools of resources, according to some embodiments. Resource management service 290 may implement a programmatic interface (e.g., API) or other interface that allows other network-based services (or a client or a provider network) to submit requests for preconfigured resources from a resource pool managed by resource management service 290. For example, a request for a cluster 930 may be received (e.g., from query tracker 340) to obtain a cluster to execute a query. Resource management service 290 may determine the appropriate pool for the request 930, a randomly (or selectively according to the techniques discussed below with regard to FIG. 14B) determine a cluster for servicing the request. Resource management service 290 may then provide the identified cluster 940 (e.g., by specifying a location, identifier, or other information for accessing the identified computing resource. Resource management service may update state information for the cluster to indicate that the cluster is leased or otherwise unavailable. Resource management service 290 may also receive requests to release a cluster 950 from a current assignment. Resource management service 290 may then update state information (e.g., the lease) for the cluster and pool to return the cluster to the pool, in some embodiments.

As indicated at 960, resource management service 290 may automatically (or in response to requests (not illustrated)), commission or decommission pool(s) of clusters 910. For example in some embodiments, resource management service 290 may perform techniques that select the number and size of computing clusters 920 for the warm cluster pool 910. The number and size of the computing clusters 920 in the warm cluster pool 910 can be determined based upon a variety of factors including, but not limited to, historical and/or expected volumes of query requests, the price of the computing resources utilized to implement the computing clusters 920, and/or other factors or considerations, in some embodiments.

Once the number and size of computing clusters 920 has been determined, the computing clusters 920 may be instantiated, such as through the use of an on-demand computing service, or virtual compute service or data processing service as discussed above in FIG. 2. The instantiated computing clusters 920 can then be configured to process queries prior to receiving the queries at the managed query service. For example, and without limitation, one or more distributed query frameworks or other query processing engines can be installed on the computing nodes in each of the computing clusters 920. As discussed above, in one particular implementation, the distributed query framework may be the open source PRESTO distributed query framework. Other distributed query frameworks can be utilized in other configurations. Additionally, distributed processing frameworks or other query engines can also be installed on the host computers in each computing cluster 920. As discussed above, the distributed processing frameworks can be utilized in a similar fashion to the distributed query frameworks. For instance, in one particular configuration, the APACHE SPARK distributed processing framework can also, or alternately, be installed on the host computers in the computing clusters 920.

Instantiated and configured computing clusters 920 that are available for use by the managed query service 270 are added to the warm cluster pool 910, in some embodiments. A determination can be made as to whether the number or size of the computing clusters 920 in the warm cluster pool needs is to be adjusted, in various embodiments. The performance of the computing clusters 920 in the warm cluster pool 910 can be monitored based on metric(s) 990 received from the cluster pool. The number of computing clusters 920 assigned to the warm cluster pool 910 and the size of each computing cluster 920 (i.e. the number of host computers in each computing cluster 920) in the warm cluster pool 910 can then be adjusted. Such techniques can be repeatedly performed in order to continually optimize the number and size of the computing clusters 920 in the warm cluster pool 910.

As indicated at 980, in some embodiments, resource management service 270 may scrub clusters(s) 980, (e.g., as a result of the lease state transitioning to expired or terminated) by causing the cluster to perform operations (e.g., a reboot, disk wipe, memory purge/dump, etc.) so that the cluster no longer retains client data and is ready to process another query. For example, resource management service 290 may determine whether a computing cluster 920 is inactive (e.g. the computing cluster 920 has not received a query in a predetermined amount of time). If resource management service 290 determines that the computing cluster 920 is inactive, then the computing cluster 920 may be disassociated from the submitter of the query. The computing cluster 920 may then be “scrubbed,” such as by removing data associated with the submitter of the queries from memory (e.g. main memory or a cache) or mass storage device (e.g. disk or solid state storage device) utilized by the host computers in the computing cluster 920. The computing cluster 920 may then be returned to the warm cluster pool 910 for use in processing other queries. In some embodiments, some clusters that are inactive might not be disassociated from certain users in certain scenarios. In these scenarios, the user may have a dedicated warm pool of clusters 910 available for their use.

As indicated at 960, in some embodiments, resource management service 290 may receive requests to configure resources or a pool of resources. For example, a request to configure a pool of resources may identify a type or size of cluster, a processing engine, machine image, or software to execute for individual clusters in the pool. In some embodiments, the request may indicate a maximum number of resources in the pool, a minimum number of idle resources in the pool, and a maximum number of idle resources in the pool. As indicated at 970, resource management service may receive a request to configure or specify a pool modification event for a pool, in some embodiments. For example, the pool modification event may be defined according to one or more criteria, such as the minimum number of idle resources, maximum number of idle resources, average job execution time thresholds, pool or resource lifecycle/state conditions, or any other set of one or more criteria that may be evaluated to detect a pool modification event.

As noted above, one or more pool modification events may be detected or otherwise triggered based on the state of a resource pool. FIG. 10 is a state diagram illustrating different resource pool states tracked by a resource manager service, according to some embodiments. Start state 1010 may be an initial state for a resource pool that has been identified or determined for creation (e.g., in response to a request to create a pool). Various different characteristics may be implemented to determine state transitions for a resource pool. In one embodiment, such characteristics may include minimum idle resource count (e.g., a minimum number of resources to keep available for clients), a maximum idle resource count (e.g., a maximum number of resources to keep available for clients), and an overall maximum resource count (e.g., a maximum number of resources that can be a member of a resource pool, whether available or leased to a client).

From start state 1010 a resource pool may either be decommissioned (e.g., in response to a request to halt or block creation of the resource pool), or transition to warmup state 1020. Warmup state 1020 may be a state where the resource pool is under provisioned, because the available or idle resource count for the pool may be less than the minimum idle resource count. Modification events to add resources to the resource pool may be triggered while the resource pool is in warm up state 1020 (e.g., either triggered only by the state being “warm up” or along with other criteria, such as criteria to throttle or limit the rate at which resources are added to a resource pool so as not to flood a resource service when creating a new pool). When the resource count for the pool equals the minimum resource count, then the resource pool may be in available state 1030, in one embodiment. Available state 1030 may be a stable state for the resource pool, neither over or under provisioned. The resource pool may remain in available state 1030 while the resource count remains between the minimum and maximum idle count. In some embodiments, if the resource count falls below minimum idle count, then the resource pool may return to warm up state 1020 (e.g., as a result of a large number of resources being leased to clients). In some embodiments, a resource pool may remain in warmup state unable to add new resources to a resource pool as adding resources may be subject to the overall count of resources (idle or leased) being less than the overall maximum count. In such a scenario, when resources are released and returned to the pool, the resource count may increase to return the resource pool to available state 1030. In scenarios where a number of resources are terminated, and removed from the pool, modification events to add resources can commence (as the overall number of resources may be less than the overall maximum).

A resource pool can transition to cool down state 1040, when the resource pool becomes over provisioned (e.g., having a current number of idle resources greater than the maximum idle count. Modification events to remove resources from the pool may be triggered (e.g., either triggered only by the state being “cool down” or along with other criteria, such as criteria to throttle or limit the rate at which resources are removed from a resource pool so as not to over compensate in the event a number of resources terminate in a short time span). A resource pool may be moved to decommission state 1050 in order to stop leasing or providing resources from the pool to clients. Leased resources in decommissioned state, in some embodiments, may be allowed to complete execution of the jobs running on the resource. Decommission state 1050 may trigger a modification events to perform decommission operations to halt, release, de-provision, or otherwise remove resources from service for a resource pool until the resource pool is empty, in one embodiment. The decommissioned pool may then transition to decommissioned state 1060, which may be logged or recorded in a resource pool history store to maintain a record of the resource pool's existence.

In addition to the life cycle of a pool triggering modification events, the lifecycle of resources may trigger modification events. Such modification events may be detected and/or performed at the resource (e.g., by an agent like managed query agent 722 in FIG. 7). FIG. 11 is a state diagram for resources implemented in a resource pool, according to some embodiments. A resource may begin in start state 1110 awaiting fulfillment. A pending resource 1120 may be a resource that has been launched but is not yet configured for processing jobs (e.g., according to a configured specified for resources in the pool, such as the query image, machine image, software applications, etc.). If an error occurs while provisioning, then the resource may be in failed state 1150, which would make the resource unable to be available to process jobs as part of the pool (and may not be counted for idle or overall resource count considerations, in some embodiments. For example, a machine image may crash or fail to load properly at one or more nodes in a cluster, in one embodiment, failing the provisioning of the resource.

For resources that are successful configured to execute jobs, the resource state may transition to ready 1130. In ready state 1130, a resource may be idle and ready to execute a job. A resource may transition out of ready state in the event of resource failure (to failed state 1150) or in the event of the resource being terminated (to terminated state 1160). Once leased or assigned to the execution of a job, resource may move to executing state 1135. Similar to ready state 1130, resource may transition out of executing state 1135 in the event of resource failure (to failed state 1150) or in the event of the resource being terminated (to terminated state 1160). Termination of a resource may, in some embodiments, occur after a time limit or other usage threshold that limits the amount of work done by a given resource. In this way, a resource that suffers from performance decline (e.g., due to age, software errors that cause memory leaks or other performance problems) or may be vulnerable to security breach can be terminated (and replaced in the pool with another resource). Upon completing execution of job, a resource may move to scrub state 1140, in some embodiments. For example, a managed query agent may detect when a cluster has completed execution of the query and report a query completion status to resource management service 290. The managed query agent may then initiate an operation to scrub the resource for reuse in the resource pool (as discussed above and below with regard to FIG. 14C). Scrubbed resources may return to resource pool by becoming in ready state 1130. In some embodiments, a scrubbed resource that fails to complete a scrub operation may move to failed state 1150 or may be terminated (e.g., due to an age/time limit for the resource).

Although FIGS. 2-11 have been described and illustrated in the context of a provider network leveraging multiple different services to implement resource management service to provide stateful management of resource pools, the various components illustrated and described in FIGS. 2-11 may be easily applied to other systems, or devices that manage pools of configured resources. As such, FIGS. 2-11 are not intended to be limiting as to other embodiments of a system that may implement stateful management of resource pools for executing jobs. FIG. 12 is a high-level flowchart illustrating various methods and techniques to implement stateful management of resources pools executing jobs, according to some embodiments. Various different systems and devices may implement the various methods and techniques described below, either singly or working together. For example, a resource management service as described above with regard to FIGS. 2-11 may implement the various methods. Alternatively, a combination of different systems and devices may implement these methods. Therefore, the above examples and or any other systems or devices referenced as performing the illustrated method, are not intended to be limiting as to other different components, modules, systems, or configurations of systems and devices.

As indicated at 1210, metrics may be obtained for pool(s) of computing resources for a network-based service that are pools of computing resources that execute jobs selectively routed by the network-based service to different ones of the computing resources, in various embodiments. Metrics may include various kinds of events or data associated with a resource pool. For example, metrics may include lifecycle events or state changes for a resource pool, as discussed above with regard to FIG. 10 or lifecycle events or state changes for a resource, as discussed above with regard to FIG. 11. Metrics may include, in some embodiments, utilization, cost, speed, or other performance or operational information for a resource (e.g., processor utilization, network bandwidth utilization, etc.). Metrics may include, in some embodiments, health metrics (e.g., failure information or states from individual resources, service or provider network infrastructure) or information based on external events, such as weather, power failures, or network events (e.g., network partitions).

As indicated at 1220, the metrics of the pools of computing resources may be evaluated to detect a modification event for at least one of the pools of the computing resources, in various embodiments. For example, aggregate values (based on individual resource metrics) may be generated (e.g., statistical values, such as averaged values, median values, etc.) and compared with one or more modification event criteria (e.g., threshold values, exact values, etc.). Modification events may be detected according to one or modification event criteria for the metrics. For example, an add resource modification event may be detected if the idle resource count is below a minimum idle threshold and if a resource addition throttle time threshold has expired.

As indicated at 1230, the at least one resource pool of computing resources may be modified according to the modification event, in various embodiments. Modification events may include events to perform any modification to the resources of a pool (e.g. cluster resize, migration, addition, removal, reconfigure, etc.) or the operation the resource pool (e.g., change the rate at which resources are added, removed, change the selection of resources to be leased, decommissioning of the resource pool, shrink the maximum size of the resource pool, increase the maximum size of the resource pool, block lease requests to the pool, etc.).

The techniques described above with regard to FIG. 12, may be implemented as part of a manually triggered pool evaluation process (e.g., by provider network or service administrator) or as part of an automated pool management process that dynamically modifies the pool according to detected modification events. FIG. 13 is a high-level flowchart illustrating techniques to monitor a resource pool for modification events according to some embodiments. As indicated at 1310, metrics for pool(s) of computing resources may be monitored for a network-based service, in some embodiments. For example, a pool monitor, such as pool monitor 820 in FIG. 8 above, may proactively evaluate the operation of a pool according to the various types of metrics discussed above. In some embodiments, metrics may be collected by polling, pinging, or otherwise requesting the metrics from different resources in the pool to generate a sample of pool performance. Different messaging or communication protocols, such as data log or event streams may be implemented so that monitoring may be performed on a live stream of metrics or on various snapshots of metrics for pool(s). Monitoring may be performed by a single worker dedicated to a single resource pool so that individual pools may not have a long lag time between the occurrence of a modification events and its detection. Monitoring of the resource pool(s) may continue as long as no modification event is detected, as indicated by the negative exit from 1320.

If a modification event is detected, then a modification may be performed corresponding to the event, as discussed above with regard to element 1230 in FIG. 12. As indicated at 1330, computing resource(s) may be added or removed for the one pool, in some embodiments.

As discussed above, resource pools can be configured to execute different types of jobs, such as queries on behalf of a managed query service. FIGS. 14A-14C describe various techniques for managing a pool of computing resources for executing queries, pools of clusters like cluster 710 in FIG. 7 discussed above. As indicated at 1402 and 1404, a number of clusters (e.g., maximum overall count) and size of computing clusters (e.g., number of hosts or nodes per cluster) for a cluster pool may be selected. As discussed above, the number and size of the computing clusters in the cluster pool can be determined based upon a variety of factors including, but not limited to, historical and/or expected volumes of query requests, the price of the computing resources utilized to implement the computing clusters, and/or other factors or considerations. As indicated at 1406, the computing clusters may be instantiated, such as through the use of an on-demand computing service, like virtual compute service 210 or processing service 220 in FIG. 2 above. The instantiated computing clusters can then be configured to process queries (e.g., by installing machine images with query engine prior to receiving queries from managed query service 270. For example, and without limitation, one or more distributed query or processing frameworks can be installed on the nodes or hosts in each of the computing clusters. As discussed above, in one embodiment, the distributed query framework may be the open source Presto distributed query framework. Other distributed query frameworks can be utilized in other embodiments. Distributed processing frameworks can also be installed on the nodes or hosts in each computing cluster, in some embodiments. Such distributed processing frameworks can be utilized in a similar fashion to distributed query frameworks. For instance, in one embodiment, the Apache Spark distributed processing framework can be installed on the host(s) or node(s) in the computing clusters.

As indicated at 1408, the instantiated and configured computing clusters that are available for use by the managed query service may be added to the cluster pool. A determination may be made, in some embodiments, as to whether the number or size of the computing clusters in the warm cluster pool should be adjusted (e.g., according to the detection of a modification event), as indicated at 1410. For example, a resource management service can monitor the performance of the computing clusters in the cluster pool to determine whether the number of size of computing clusters should be adjusted (e.g., according to the pool lifecycle discussed above with regard to FIG. 10). The number of computing clusters assigned to the cluster pool and the size of each computing cluster (i.e. the number of nodes or hosts in each computing cluster) in the cluster pool can then be adjusted. Such a technique can be repeated in order to continually optimize the number and size of the computing clusters in the cluster pool.

FIG. 14B illustrates a high-level flowchart for a technique to select a resource from a resource pool, according to some embodiments. For example, as indicated at 1452 the arrival of a query may occur (e.g., at managed query service). A request may be made to provision a cluster from a cluster pool for processing the query (e.g., to a resource management service 290 or other component that manages the cluster pool). As indicated at 1454, a computing cluster from the computing clusters in the cluster for executing the received query may be made, in various embodiments. The selection of the computing cluster for executing the query 108 can be based upon a number of factors including, but not limited to, previous queries submitted by the same requestor, desired query performance, user preferences, the amount of data to be queried, column statistics, empirical data, the price of the computing resources utilized to perform the query, other types of statistics relating to the performance of the clusters, and others, in some embodiments.

As indicated at 1456, the selected computing cluster may be removed from the cluster pool. For example, the state of the resource may be marked as leased, in one embodiment. The selected computing cluster may be associated with the submitter of the query (e.g., according to a username, identifier, or other credential linked to a user or account), as indicated at 1458. For instance, the cluster can be associated with the user when the user logs into a management console. As indicated at 1460, the query may be routed to the selected selected computing cluster 106A for execution of the query with respect to data identified in the query. Subsequent queries received from the same submitter may be routed to the selected computing cluster, in some embodiments.

FIG. 14C illustrates a technique for scrubbing a computing resource in a resource pool, according to some embodiments. As indicated at 1472, a determination may be made as to whether a computing cluster is inactive (e.g. the computing cluster has not received a query in a predetermined amount of time). If the computing cluster is determined to be inactive, then the computing cluster may be disassociated from the submitter of the query, as indicated at 1474. As indicated at 1476, the computing cluster may be “scrubbed,” in various embodiments (e.g., by removing data associated with the submitter of the queries from memory, such as a main memory or a cache, or mass storage device, disk or solid state storage device, utilized by the nodes or host computers in the computing cluster. As indicated at 1478, the computing cluster may be returned to the cluster pool for use in processing other queries. In some embodiments, it may be that clusters may not be disassociated from certain users in certain scenarios. In these scenarios, the user may have a dedicated pool of clusters available for their use.

The methods described herein may in various embodiments be implemented by any combination of hardware and software. For example, in one embodiment, the methods may be implemented by a computer system (e.g., a computer system as in FIG. 17) that includes one or more processors executing program instructions stored on a computer-readable storage medium coupled to the processors. The program instructions may be configured to implement the functionality described herein (e.g., the functionality of various servers and other components that implement the network-based virtual computing resource provider described herein). The various methods as illustrated in the figures and described herein represent example embodiments of methods. The order of any method may be changed, and various elements may be added, reordered, combined, omitted, modified, etc.

FIG. 15 is a logical block diagram that shows an illustrative operating environment that includes a service provider network that can implement aspects of the functionality described herein, according to some embodiments. As discussed above, the service provider network 200 can provide computing resources, like VM instances and storage, on a permanent or an as-needed basis. Among other types of functionality, the computing resources provided by the service provider network 200 can be utilized to implement the various services described above. As also discussed above, the computing resources provided by the service provider network 200 can include various types of computing resources, such as data processing resources like VM instances, data storage resources, networking resources, data communication resources, network services, and the like.

Each type of computing resource provided by the service provider network 200 can be general-purpose or can be available in a number of specific configurations. For example, data processing resources can be available as physical computers or VM instances in a number of different configurations. The VM instances can execute applications, including web servers, application servers, media servers, database servers, some or all of the services described above, and/or other types of programs. The VM instances can also be configured into computing clusters in the manner described above. Data storage resources can include file storage devices, block storage devices, and the like. The service provider network 200 can also provide other types of computing resources not mentioned specifically herein.

The computing resources provided by the service provider network maybe implemented, in some embodiments, by one or more data centers 1304A-1304N (which might be referred to herein singularly as “a data center 1304” or in the plural as “the data centers 1304”). The data centers 1304 are facilities utilized to house and operate computer systems and associated components. The data centers 1304 typically include redundant and backup power, communications, cooling, and security systems. The data centers 1304 can also be located in geographically disparate locations. One illustrative configuration for a data center 1304 that can be utilized to implement the technologies disclosed herein will be described below with regard to FIG. 16.

The customers and other users of the service provider network 200 can access the computing resources provided by the service provider network 200 over a network 1302, which can be a wide area communication network (“WAN”), such as the Internet, an intranet or an Internet service provider (“ISP”) network or a combination of such networks. For example, and without limitation, a computing device 1300 operated by a customer or other user of the service provider network 200 can be utilized to access the service provider network 200 by way of the network 1302. It should be appreciated that a local-area network (“LAN”), the Internet, or any other networking topology known in the art that connects the data centers 1304 to remote customers and other users can be utilized. It should also be appreciated that combinations of such networks can also be utilized.

FIG. 16 is a logical block diagram illustrating a configuration for a data center that can be utilized to implement aspects of the technologies disclosed herein, according to various embodiments. is a computing system diagram that illustrates one configuration for a data center 1304 that implements aspects of the technologies disclosed herein for providing managed query execution, such as managed query execution service 270, in some embodiments. The example data center 1304 shown in FIG. 16 includes several server computers 1402A-1402F (which might be referred to herein singularly as “a server computer 1402” or in the plural as “the server computers 1402”) for providing computing resources 1404A-1404E.

The server computers 1402 can be standard tower, rack-mount, or blade server computers configured appropriately for providing the computing resources described herein (illustrated in FIG. 16 as the computing resources 1404A-1404E). As mentioned above, the computing resources provided by the provider network 200 can be data processing resources such as VM instances or hardware computing systems, computing clusters, data storage resources, database resources, networking resources, and others. Some of the servers 1402 can also execute a resource manager 1406 capable of instantiating and/or managing the computing resources. In the case of VM instances, for example, the resource manager 1406 can be a hypervisor or another type of program may enable the execution of multiple VM instances on a single server computer 1402. Server computers 1402 in the data center 1304 can also provide network services and other types of services, some of which are described in detail above with regard to FIG. 2.

The data center 1304 shown in FIG. 16 also includes a server computer 1402F that can execute some or all of the software components described above. For example, and without limitation, the server computer 1402F can execute various components for providing different services of a provider network 200, such as the managed query service 270, the data catalog service 280, resource management service 290, and other services 1410 (e.g., discussed above) and/or the other software components described above. The server computer 1402F can also execute other components and/or to store data for providing some or all of the functionality described herein. In this regard, it should be appreciated that the services illustrated in FIG. 16 as executing on the server computer 1402F can execute on many other physical or virtual servers in the data centers 1304 in various configurations.

In the example data center 1304 shown in FIG. 16, an appropriate LAN 1406 is also utilized to interconnect the server computers 1402A-1402F. The LAN 1406 is also connected to the network 1302 illustrated in FIG. 15. It should be appreciated that the configuration and network topology described herein has been greatly simplified and that many more computing systems, software components, networks, and networking devices can be utilized to interconnect the various computing systems disclosed herein and to provide the functionality described above. Appropriate load balancing devices or other types of network infrastructure components can also be utilized for balancing a load between each of the data centers 1304A-1304N, between each of the server computers 1402A-1402F in each data center 1304, and, potentially, between computing resources in each of the data centers 1304. It should be appreciated that the configuration of the data center 1304 described with reference to FIG. 16 is merely illustrative and that other implementations can be utilized.

Embodiments of a managed query execution as described herein may be executed on one or more computer systems, which may interact with various other devices. One such computer system is illustrated by FIG. 17. In different embodiments, computer system 2000 may be any of various types of devices, including, but not limited to, a personal computer system, desktop computer, laptop, notebook, or netbook computer, mainframe computer system, handheld computer, workstation, network computer, a camera, a set top box, a mobile device, a consumer device, video game console, handheld video game device, application server, storage device, a peripheral device such as a switch, modem, router, or in general any type of computing device, computing node, compute node, computing system compute system, or electronic device.

In the illustrated embodiment, computer system 2000 includes one or more processors 2010 coupled to a system memory 2020 via an input/output (I/O) interface 2030. Computer system 2000 further includes a network interface 2040 coupled to I/O interface 2030, and one or more input/output devices 2050, such as cursor control device 2060, keyboard 2070, and display(s) 2080. Display(s) 2080 may include standard computer monitor(s) and/or other display systems, technologies or devices. In at least some implementations, the input/output devices 2050 may also include a touch- or multi-touch enabled device such as a pad or tablet via which a user enters input via a stylus-type device and/or one or more digits. In some embodiments, it is contemplated that embodiments may be implemented using a single instance of computer system 2000, while in other embodiments multiple such systems, or multiple nodes making up computer system 2000, may host different portions or instances of embodiments. For example, in one embodiment some elements may be implemented via one or more nodes of computer system 2000 that are distinct from those nodes implementing other elements.

In various embodiments, computer system 2000 may be a uniprocessor system including one processor 2010, or a multiprocessor system including several processors 2010 (e.g., two, four, eight, or another suitable number). Processors 2010 may be any suitable processor capable of executing instructions. For example, in various embodiments, processors 2010 may be general-purpose or embedded processors implementing any of a variety of instruction set architectures (ISAs), such as the x86, PowerPC, SPARC, or MIPS ISAs, or any other suitable ISA. In multiprocessor systems, each of processors 2010 may commonly, but not necessarily, implement the same ISA.

In some embodiments, at least one processor 2010 may be a graphics processing unit. A graphics processing unit or GPU may be considered a dedicated graphics-rendering device for a personal computer, workstation, game console or other computing or electronic device. Modern GPUs may be very efficient at manipulating and displaying computer graphics, and their highly parallel structure may make them more effective than typical CPUs for a range of complex graphical algorithms. For example, a graphics processor may implement a number of graphics primitive operations in a way that makes executing them much faster than drawing directly to the screen with a host central processing unit (CPU). In various embodiments, graphics rendering may, at least in part, be implemented by program instructions configured for execution on one of, or parallel execution on two or more of, such GPUs. The GPU(s) may implement one or more application programmer interfaces (APIs) that permit programmers to invoke the functionality of the GPU(s). Suitable GPUs may be commercially available from vendors such as NVIDIA Corporation, ATI Technologies (AMD), and others.

System memory 2020 may store program instructions and/or data accessible by processor 2010. In various embodiments, system memory 2020 may be implemented using any suitable memory technology, such as static random access memory (SRAM), synchronous dynamic RAM (SDRAM), nonvolatile/Flash-type memory, or any other type of memory. In the illustrated embodiment, program instructions and data implementing desired functions, such as those described above are shown stored within system memory 2020 as program instructions 2025 and data storage 2035, respectively. In other embodiments, program instructions and/or data may be received, sent or stored upon different types of computer-accessible media or on similar media separate from system memory 2020 or computer system 2000. Generally speaking, a non-transitory, computer-readable storage medium may include storage media or memory media such as magnetic or optical media, e.g., disk or CD/DVD-ROM coupled to computer system 2000 via I/O interface 2030. Program instructions and data stored via a computer-readable medium may be transmitted by transmission media or signals such as electrical, electromagnetic, or digital signals, which may be conveyed via a communication medium such as a network and/or a wireless link, such as may be implemented via network interface 2040.

In one embodiment, I/O interface 2030 may coordinate I/O traffic between processor 2010, system memory 2020, and any peripheral devices in the device, including network interface 2040 or other peripheral interfaces, such as input/output devices 2050. In some embodiments, I/O interface 2030 may perform any necessary protocol, timing or other data transformations to convert data signals from one component (e.g., system memory 2020) into a format suitable for use by another component (e.g., processor 2010). In some embodiments, I/O interface 2030 may include support for devices attached through various types of peripheral buses, such as a variant of the Peripheral Component Interconnect (PCI) bus standard or the Universal Serial Bus (USB) standard, for example. In some embodiments, the function of I/O interface 2030 may be split into two or more separate components, such as a north bridge and a south bridge, for example. In addition, in some embodiments some or all of the functionality of I/O interface 2030, such as an interface to system memory 2020, may be incorporated directly into processor 2010.

Network interface 2040 may allow data to be exchanged between computer system 2000 and other devices attached to a network, such as other computer systems, or between nodes of computer system 2000. In various embodiments, network interface 2040 may support communication via wired or wireless general data networks, such as any suitable type of Ethernet network, for example; via telecommunications/telephony networks such as analog voice networks or digital fiber communications networks; via storage area networks such as Fibre Channel SANs, or via any other suitable type of network and/or protocol.

Input/output devices 2050 may, in some embodiments, include one or more display terminals, keyboards, keypads, touchpads, scanning devices, voice or optical recognition devices, or any other devices suitable for entering or retrieving data by one or more computer system 2000. Multiple input/output devices 2050 may be present in computer system 2000 or may be distributed on various nodes of computer system 2000. In some embodiments, similar input/output devices may be separate from computer system 2000 and may interact with one or more nodes of computer system 2000 through a wired or wireless connection, such as over network interface 2040.

As shown in FIG. 17, memory 2020 may include program instructions 2025, may implement the various methods and techniques as described herein, and data storage 2035, comprising various data accessible by program instructions 2025. In one embodiment, program instructions 2025 may include software elements of embodiments as described herein and as illustrated in the Figures. Data storage 2035 may include data that may be used in embodiments. In other embodiments, other or different software elements and data may be included.

Those skilled in the art will appreciate that computer system 2000 is merely illustrative and is not intended to limit the scope of the techniques as described herein. In particular, the computer system and devices may include any combination of hardware or software that can perform the indicated functions, including a computer, personal computer system, desktop computer, laptop, notebook, or netbook computer, mainframe computer system, handheld computer, workstation, network computer, a camera, a set top box, a mobile device, network device, internet appliance, PDA, wireless phones, pagers, a consumer device, video game console, handheld video game device, application server, storage device, a peripheral device such as a switch, modem, router, or in general any type of computing or electronic device. Computer system 2000 may also be connected to other devices that are not illustrated, or instead may operate as a stand-alone system. In addition, the functionality provided by the illustrated components may in some embodiments be combined in fewer components or distributed in additional components. Similarly, in some embodiments, the functionality of some of the illustrated components may not be provided and/or other additional functionality may be available.

Those skilled in the art will also appreciate that, while various items are illustrated as being stored in memory or on storage while being used, these items or portions of them may be transferred between memory and other storage devices for purposes of memory management and data integrity. Alternatively, in other embodiments some or all of the software components may execute in memory on another device and communicate with the illustrated computer system via inter-computer communication. Some or all of the system components or data structures may also be stored (e.g., as instructions or structured data) on a computer-accessible medium or a portable article to be read by an appropriate drive, various examples of which are described above. In some embodiments, instructions stored on a non-transitory, computer-accessible medium separate from computer system 2000 may be transmitted to computer system 2000 via transmission media or signals such as electrical, electromagnetic, or digital signals, conveyed via a communication medium such as a network and/or a wireless link. Various embodiments may further include receiving, sending or storing instructions and/or data implemented in accordance with the foregoing description upon a computer-accessible medium. Accordingly, the present invention may be practiced with other computer system configurations.

It is noted that any of the distributed system embodiments described herein, or any of their components, may be implemented as one or more web services. For example, leader nodes within a data warehouse system may present data storage services and/or database services to clients as network-based services. In some embodiments, a network-based service may be implemented by a software and/or hardware system designed to support interoperable machine-to-machine interaction over a network. A network-based service may have an interface described in a machine-processable format, such as the Web Services Description Language (WSDL). Other systems may interact with the web service in a manner prescribed by the description of the network-based service's interface. For example, the network-based service may define various operations that other systems may invoke, and may define a particular application programming interface (API) to which other systems may be expected to conform when requesting the various operations.

In various embodiments, a network-based service may be requested or invoked through the use of a message that includes parameters and/or data associated with the network-based services request. Such a message may be formatted according to a particular markup language such as Extensible Markup Language (XML), and/or may be encapsulated using a protocol such as Simple Object Access Protocol (SOAP). To perform a web services request, a network-based services client may assemble a message including the request and convey the message to an addressable endpoint (e.g., a Uniform Resource Locator (URL)) corresponding to the web service, using an Internet-based application layer transfer protocol such as Hypertext Transfer Protocol (HTTP).

In some embodiments, web services may be implemented using Representational State Transfer (“RESTful”) techniques rather than message-based techniques. For example, a web service implemented according to a RESTful technique may be invoked through parameters included within an HTTP method such as PUT, GET, or DELETE, rather than encapsulated within a SOAP message.

The various methods as illustrated in the FIGS. and described herein represent example embodiments of methods. The methods may be implemented in software, hardware, or a combination thereof. The order of method may be changed, and various elements may be added, reordered, combined, omitted, modified, etc.

Various modifications and changes may be made as would be obvious to a person skilled in the art having the benefit of this disclosure. It is intended that the invention embrace all such modifications and changes and, accordingly, the above description to be regarded in an illustrative rather than a restrictive sense.

STATEFUL RESOURCE POOL MANAGEMENT FOR JOB EXECUTION

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

International Classifications

Abstract

Description

Claims

RELATED APPLICATIONS

Provisional Applications (1)