Patent Application
20160072683
This disclosure relates to a complex system architecture and analytics engine for building, maintaining, and analyzing infrastructure layouts. This disclosure also relates to complex system architecture for determination of policies and execution designs based on the infrastructure layouts.
Described below is a context-aware architecture that maps northbound services to southbound service fulfillment models. In various implementations, northbound services include customized service offerings to an end user. For example, northbound services may include data connectivity for an automobile, messaging capability on a phone, remote storage for a mobile device, or virtually any end user deliverable service. The context-aware architecture may further be used for workflow generation to support the deployment of the northbound services or completion of other tasks. For example, the workflow may include ordered scripts that may the implementation of intervening components connecting the northbound service to fulfillment via the southbound service. Additionally or alternatively, the context aware architecture may facilitate policy enforcement. The capture of relational data among the component providing the northbound and southbound services may allow comparison to policy models. For example, when the captured relational data differs from the policy model, the system may generate alerts and/or otherwise handle the policy deviation.
In various implementations, southbound services include service capabilities to support fulfillment of the northbound services. For example, southbound service may include WiFi connectivity, text messaging though a particular service provider, web storage services, and/or other fulfillment services.
In some cases, it may be advantageous to implement a context-aware infrastructure layout architecture to map northbound offered services to southbound service fulfillment models. For example, the context-aware infrastructure layout architecture may be applied in services, such as connectivity for cars and/or homes fulfilled though telecommunications and/or internet service providers (ISPs); cloud brokers serving software as a service (SaaS), platform as a service (PaaS), and/or infrastructure as a service (IaaS) fulfilled through Amazon web services (AWS), Cloud Foundry, Open Stack based data centers, and/or other web service providers; digital media content platforms serving content and/or advertising to digital media customers fulfilled by content providers and social media channels; and/or other northbound offerings mapped to southbound fulfillment models.
In some cases the context-aware infrastructure layout architecture may allow for the customized use of southbound services, and such services may be implemented in a generic or generalized fashion. For example, a northbound service may be mapped to a portion of capabilities provided by one or more southbound services. Thus, the south bound service may not necessarily be customized to the particular northbound service. Rather, fulfillment for the northbound service may arise through a request and execution of one or more capabilities of the southbound services. As the number of northbound use cases grows, the context-aware architecture allows for accelerated configuration and deployment of southbound services and then assists in the management of updates.
In various implementations, a core layout may describe the allowable relationships within scope of the platform. In some cases, the allowable relationships may include mobile managed services (MMS).
Extensions to the core layout may broaden the scope of relationships defined in the core layout through appending new domain layouts and mapping how the new domain layouts interrelate with the core layout and/or previous extensions. Various instances of northbound services being mapped to southbound fulfillment models may be implemented via the core layout and available extensions.
In various implementations, the core layout may act as a central point for connecting northbound and southbound services. The core layout may facilitate determination of how the usage cases described by the northbound services are fulfilled by the southbound services. In some cases a user interface (e.g., in the form of an online website or mobile application) may be used to guide a user through the configuration of an instance of the core layout and/or extensions. The process may generate an instance that maps a northbound service to fulfillment via one or more southbound services.
The configuration logic 300 may implement discovery of effects introduced by the changes to the core layout through the extensions for the developed workflow (312). The configuration logic 300 may traverse the model (e.g., automatically or upon operator request) to find instances that have been deployed under other components that share the same parent (314). The logic 300 notify the operator of the change (316).
Turing now to
In various implementations for various northbound services, the context-aware layout architecture may obviate some redesigns of data and services to support data and services for different types of interactions. Various northbound services may leverage the same underlying data, but the context for the data may be changed depending on the particular northbound service accessing the data. In the example scenario 1700, the owner, manufacturer, and insurer may use the same data resource, car connectivity, but for different contexts.
Various implementations may employ layout logic 700, discussed below, to map the northbound service offerings through the core platform model to the available southbound fulfillment services. In some cases, the layout logic 700 may output the resulting mappings of how southbound services fulfill offered northbound services in a manifest. For example, a manifest may include a file that lists southbound services and accompanying configurations used to support a particular northbound service.
The layout logic 700 may access a core layout (701). The layout logic 700 may determine a northbound service (702). The layout logic 700 may determine one or more services supporting the northbound service (704). The layout logic 700 may determine whether a southbound service is available for fulfilment of a selected service of the supporting services (706). If a southbound service is available, the layout logic 700 may a path from the selected service to the southbound service to generate an instance (708). Via the traversal, the layout logic 700 may create a set of nodes corresponding to the traversed nodes and edges to generate the instance (710).
If a southbound service is not available, the layout logic 700 may determine whether an extension may provide support for the unavailable southbound service (712). The layout logic 700 may adjust nodes in the core layout to support joining with the extension (714). The layout logic 700 may add nodes to the core layout to support joining with the extension (716). The layout logic 700 may remove nodes in the core layout to support joining with the extension (718). The layout logic 700 may integrate one or more nodes from the extension into the core layout (720). The layout logic 700 may replace the core layout with the joined core layout and extension (722). Once the extension is integrated, the layout logic 700 may proceed to traverse the newly formed path (708).
In some implementations, the core layout, extensions, northbound services, southbound services, and/or other aspects of the layout architecture may be described using a web resource layout platform. For example, a web resource layout platform may include a resource description framework (RDF) and layout logic 700 to capture and manage the fulfillment of the northbound services via the southbound services.
In various implementations, the southbound services may be accessed through one or more application programming interfaces (APIs). In some cases, the APIs may allow for adjustment of parameters for execution of the southbound services. Thus, northbound instances may be implemented via one or more API access based requests to the available southbound services.
In an example scenario, the infrastructure layout architecture may be used to support deployment of devices and services to implement Internet of Things (IoT) based services. For example, the IoT based services may include automobile connectivity, home appliance connectivity, health service connectivity, and/or other Internet integration services.
In some implementations, the core layout may include multiple cascading components. For example, an application platform layout may be cascaded over an infrastructure layout.
In an example cascaded scenario, the layout logic 700 may determine one or more application services and accompanying parameters to fulfill a northbound service. For example, the northbound service may include an organization specific service. Based on an analysis of northbound service for an application platform layout, the southbound fulfillment services may be identified and requested by the layout logic 700. In some cases, the northbound services of an infrastructure platform layout may supply the southbound service fulfillment for the application platform layout.
Layout logic 700 may determine the one or more infrastructure services and accompanying parameters to fulfill the southbound application services from the cascaded application platform layout. The infrastructure layout may include software code repositories, compute commands, scripts, variables, and/or other components that can be used to populate a deployable manifest.
The cascaded approach may be extended beyond two cascaded layouts. For example, another cascaded layout, e.g., describing details of cloud functionality and/or of a sensor platform, can be cascaded below the infrastructure layout. Additionally or alternatively a cascaded layout may be added above the northbound models describing the organization specific service.
In various implementations, the various cascaded layouts may be implemented with their own independent layout logic 700 for organization and execution. However, in some implementations, multiple cascaded layers may be controlled by unified layout logic 700. Additionally or alternatively, distinct, interoperable layout logic 700 instances may be used to control the cascaded layers.
As discussed above, the core layout, northbound services, and southbound services may be represented in a web resource layout. Table 1 shows an example RDF-based layout that captures a triple including a subject, predicate (e.g., relationship), and object. The triple describes how the subject relates to the object. In the example RDF-based layout, the project is a northbound service and the application platform is a southbound service supporting the northbound service.
The RDF format differs from relational database tables whose relations are pre-defined at design time and are implicit across the rows and columns of a table. Instead with RDF, relationships are explicitly stored as properties. In a graph based representation, these properties may be associated with the edges that connect vertices in the graph. The explicit storage of these relationships provides the context for interpretation of the parameters for configuration of the service or product. Further, storage of the relationship in addition to the parameter allows for alteration of the relationship without altering the parameter and vice versa. The independent adjustment of these factors allows the layout logic 700 to support extensions or changes to northbound services via adjustment of relationships or parameters rather than using a single degree of freedom.
The following pseudocode is an example SPARQL protocol and RDF query language (SPARQL) implementation to support verification of the example rule above:
The following pseudocode is example SPARQL rule implementation to support configuration consistent with the example rule above:
To on-board a new northbound or southbound service to a core layout (portion S 230), the layout logic 700 may generate a new graph portion T 270 that includes subclasses and types from a sub graph of the core layout S 230.
For node vt from the portion S, a new node vt′ in T 270 that inherits attributes, properties, and rules associated from vt in S 230. In some cases, an instance may map to a subset of the core layout when full capability is not used or fulfilled, for example, when a vendor product implements a specific subset of the capabilities of a service provision system.
In some implementations, T 270 may also capture additional attributes, properties, and vertices and edges not included S 230. For example, T 270 may rely on portions of the extended core layout E 250. Additionally or alternatively, T 270 may remove attributes, properties, vertices, and edges in S 230.
The layout logic 700 may check that vt′ adheres to inherited rules. The layout logic 700 may also generate indicators of exceptions where vt′ in T 270 does not adhere to a rule for vt. The layout logic 700 may then relate vt′ to vt by applying a property. For example, the nodes may be related using the property: ‘subclass_of’ and/or another relationship. This relationship may be generalized to connect T 270 and S 230. Additionally, the relationship generation may be repeated for nodes v′ within T. The layout logic 700 may then replace S 230 as the core layout with the connected T 270 and S 230 layout to accommodate the on-boarded northbound or southbound service.
Additionally or alternatively, an input to the layout logic 700 may specify an instance T 270 of the core layout S 230, where T 270 captures a specific configuration of a subset of S. For example, the input may be supplied from a user interface, such as a wizard application, that facilitates navigation from an initial subset of vertices v in S 230 and queries the user about which vertices include and how to configure the vertices and edges in the selected subset. The subset of S 230 that is presented to the user may depend on the initial location of the vertex in S 230 and the rules to the attached vertex and the edges.
Once the input indicates the inclusion of a node v from S, the layout logic 700 may create a vertex v′ in T 270 that inherits attributes, properties, and rules associated v in S 230. In some cases, the layout logic 700 may attach instance specific properties to selected vertices and edges.
As discussed above, T 270 may include elements, such as attributes, properties, and vertices and edges, not in S 230 or may omit elements present in S 230. The layout logic 700 may check that vertices adhere to inherited rules. The layout logic 700 may also generate indicators of exceptions where vertices in T 270 do not adhere to one or more rules form S 230.
Additionally or alternatively, a template may be used to identify a subset of S 230 that forms T.
Northbound services may be treated as fulfilled where one or more vertices in the core layout representing the northbound service are connected, via other vertices or edges, to one or more vertices representing southbound services.
Additionally or alternatively, the layout logic 700 may update the extended layout E 250. For example, an operation may include the layout logic 700 deleting or adjusting a vertex or an edge in G 240. For example, a vertex v in E 250 may be transformed to vertex w through property adjustments. In the case of instance Y 280 that overlaps with extension G 240, the layout logic 700 may record a corresponding transformation of w to w″ within Y 280. For deletions or adjustments to edges within E, the corresponding edges in the Y 280 may be updated.
Additionally or alternatively, the layout logic 700 may add vertexes in E 250. In some implementations, the addition of a vertex in E 250 may be handled similarly to the addition of an edge that is considered linked through E with a relationship. For example, such an addition may include a vertex x that is either a source that connects to E or a destination that is connected to a source vertex in E.
In some implementations, a list L of one or more instances may be maintained by the layout logic 700. Upon detection that such a change impacts a list of instances L, the layout logic 700 main send a trigger to the instances to initiate regeneration of deployment artifacts. In other words, the layout logic 700 may determine if the changes alter interactions between northbound and southbound services. To avoid improperly linked services and/or out of date deployments, the instances are regenerated to such that the instances are in compliance with the adjusted, added, or deleted elements in the instance.
In various implementations, a fulfilled northbound service may be stored as a template. In some cases, the template may include a text representation of the resources and services on which the northbound service depends. For example, the text representation may include a YAML ain't markup language (YAML) file, a JavaScript object notation (JSON) file, and/or an extensible markup language (XML) file. The template may include indications of resources and/or dependencies for creating, maintaining, deleting, and/or performing other lifecycle operations on the northbound service stack. In some cases, the template may instantiate the resources and dependencies by assigning stored property values.
In some cases, cloud providers may have individualized application management systems. Therefore, the actions used to configure, generate and execute a workflow including application infrastructure lifecycle events, such as, installations, application configurations, starts, stops, deletions or other application infrastructure lifecycle events, may differ from provider to provider. For example, when deploying a MySQL database service, one may first create a virtual machine, configure the Operating System, install and configure the MySQL server, populate the database, and then grant the selected permissions to selected users. However, this process may vary for other database services. When switching between the individualized application management systems, processes generated for a first system may not necessarily be immediately portable to a second system. For example, AWS OpsWorks uses actions definitions that are specific to the AWS cloud.
Some cloud automation tools, such as, RightScale, vFabric Application Director have been developed to reduce the friction when migrating applications from one cloud to another. However, these automation tools may rely on proprietary models. Some standardized cloud models have been proposed to model the lifecycle of the cloud applications, such as topology and orchestration specification for cloud applications (TOSCA). However, these tools may not necessarily provide a systematic way to extend this standardized model to apply global optimization policies or hierarchical based policies during workflow generation or to verify the cloud service along its lifecycle.
In various implementations, the deployment and/or management of cloud services requires orchestration of lifecycle events/operations of multiple components. Because the events/operations of different components can be generated and/or executed in varied orders, unanticipated interactions among the components may occur. In some cases, these unanticipated interactions may lead to failures. The large number of interactions likely to occur coupled with the wide variety of cloud capabilities presents challenges in generating feasible workflows that execute as expected.
The space of possible interactions in a given information technology (IT) infrastructure domain may be large. For example, there are many options for provisioning a DBMS server and populating a database on the server. A first implementation may provision a SQLServer database server on Windows Operating System using an AWS elastic compute cloud (EC2) instance. Additionally or alternatively, an implementation may use a MySQL database server on the Ubuntu Operating System using the Google Compute Engine (GCE). Individual components for these examples may create a large number of dependencies. Further, the components may also have different requirements; preferences; and policies for different users, organizations and/or environments. Exhaustively enumerating the workflows that satisfy these may be a challenge.
Executing individual events in a workflow may result in differing states depending on the event. In some cases, tests may be performed on the individual possible states when changes are made. However, because of the large number of execution order permutations, performing these tests may consume significant resources.
Test oracles are mechanisms that determine whether a workflow has been executed correctly. In some cases, test oracles may interleave their verifications with event/operations because an incorrect state may prevent further execution of the process. For example, an event in a process may not be allowed if an incorrect state results at the time of execution for the event. Thus, execution of the workflow may be terminated as soon as an error is detected to preserve resources. Additionally or alternatively, interleaved verification may be used to facilitate error identification. For example, interleaved verification may allow a system to determine a faulty event based on the detection of an erroneous state following the execution of the event.
When the northbound service is fulfilled, the flow generation logic 800 may traverse the path from the source node to a destination node representing a fulfilling southbound service (816). The flow generation logic may generate a workflow based on the components, dependencies, and/or relationships traced along the traversed path (818). The flow generation logic 800 may determine whether multiple southbound services provide fulfilment for the selected source node (820). For example, a northbound service may depend on multiple southbound services and/or redundantly fulfilled by multiple southbound services. When multiple southbound services provide fulfilment, the flow generation logic may repeat the traversal for the remaining southbound services and corresponding destination nodes (816). Once the flow generation logic 800 has completed tracing the paths to the destination nodes, the flow generation logic may output the workflows (822). For example, the workflows may by ordered scripts to support deployment of the northbound service.
In various implementations, workflow generation based on the context-aware infrastructure layout architecture may facilitate predictable execution of the workflows across differencing cloud environments. The core layout may provide a hierarchical graph-based model of a service provision system. In some cases flow generation logic may traverse the core layout and enumerate paths that satisfy various service components, such as northbound or southbound services. In an example implementation, an RDF system may be used to represent the TOSCA model. The combination may facilitate modelling of IT infrastructure components, deployment artifacts and lifecycle scripts, and the relationships among the components, artifacts, and scripts. In some implementations, roles may be supported. The roles may be used to model different users, organizations and environments. The roles may be used to create a query library that containing queries to enforce dependencies, preferences, and policies. A query may receive an input and return results that comply with a policy that query is enforcing. For example, to enforce a cost minimization policy, an operator may select a query that corresponds to a cost minimization policy from the query library and then may supply the minimum CPU/Disk/RAM value as an input to the selected query. The query may return a result that satisfies the policy.
A TOSCA RDF graph model may contain different types of nodes and relationships. We can identify one or more specific nodes or relationships that satisfy certain preferences by writing queries, for example, SPARQL queries. For example, to implement a policy asserts a preference for the cheapest virtual machine (VM) instances from available cloud providers, the pseudocode below may be used, the result will be the provider's name, e.g., AWS EC2, GCE; the instance type, e.g., m3.medium, f1-micro; and the price:
To refine the example query to enforce particular preferences, e.g., minimum RAM, or other preferences, the pseudocode shown below may be used. The input to the code may be the MINIMUM_RAM field, and output will be the same as that of the previous pseudocode. In some cases, the pseudocode above may produce that same output as the pseudocode below with the MINIMUM_RAM input set to 0:
A query library may include a bundle of queries, with individual names, inputs and outputs. Users may select queries from the library based on names or other references to enforce particular policies. Some operators may edit and/or extend the library with refined and/or new queries based on users' requests.
In various implementations, the input to the flow generation logic is the core layout and a source node for a path. The output of the flow generation logic may include workflows that include ordered scripts, such as scripts that will be run on a provisioned virtual machine.
In an example implementation, a TOSCA meta-model may be used for structuring and managing IT services. In the example implementation, a deployed service is an instance of a service template, which is derived by instantiating the topology template and orchestrated via a specified plan. The topology template may be used to define the structure of a service. For example the core layout discussed above may be used as a topology template. The node and relationship types in a layout may define the properties associated with that type and lifecycle operations (via interfaces) available for invocation when a service template is instantiated. For example, a MySQL node type may represent a MySQL database management system (DBMS)), which has properties, such as, root_password, dbms_port; and life cycle operations such as install, configure, start, stop. These properties and life cycle operations may map to actual implementation artifacts, such as, scripts, Chef recipes, and/or other configuration definition elements. Chef is a configuration management tool available from Opscode, and may provide a fundamental configuration element that defines what may be needed to configure part of a system, e.g., install a MySQL server. Additionally or alternatively, deployment artifacts for a node type, such as MySQL RPM bundle for installation, may be present. In particular, a node type may be annotated with requirements and capabilities, such that a relationship can be built between a node using a specific capability and another node providing the specific capability. In some cases, an analysis of the capabilities used by a node and the capabilities offered by other node may identify opportunities for substitution.
In the TOSCA-based example implementation, the resources are components, e.g., Operating System, Virtual Machine, and/or other components, in the infrastructure domain; properties of the components, e.g., IP address, port number, and/or other properties; and artifacts, e.g., deployment scripts, configuration files, archived installables, and/or other artifacts. Using the RDF systems a statements about resources may be made by the layout logic 700 as triples: <Subject> <Predicate> <Object>. The layout logic 700 may group the statements into graphs.
In the TOSCA-based example implementation, nodes represent resources (subjects and objects), and the edges 1102 between nodes 1104 represent relationships (predicates). The upper half of the graph 1150 defines the infrastructure at the “type” level. In other words, the upper half of the graph shows what types of components are present the relationships the components have. The bottom half 1170 in the graph defines the actual instances of particular types within the infrastructure. For example, the green line in the graph represents <is hosted on> type of relationship; the pink line represents <is an instance of> type of relationship; and the navy line represents <is a subclass of> type of relationship. For example triples using the format may include: <SugarCRM—1> <is an instance of> <SugarCRM>, <SugarCRM> <is a subclass of> <WebApp>, <WebApp> <is hosted on> <WebServer>.
The system may define, for example, the following types of relationship in the graph:
<has property>—modifies a component that has a property. For this type, the object may be satisfied by a property node. For example, <virtual machine> <has property> <ip address>, <software> <has property> <version>, <EC2 Instance> <has property> <access id>.
<has artifact>—modifies a component that has a property. For this type, the object may be satisfied by an artifact node. The artifact may be in the form of a script, a configuration file, archived installables, or other artifact. Additionally or alternatively, different scripts may serve different purposes, e.g., various lifecycle events such as install, delete, start, stop, configure may have different purposes.
<is subclass of>—modifies a node is a subclass of another node.
<is instance of>—modifies a node, e.g., a node in the bottom half of the graph, that is an instance of another node, e.g., a node in the upper half.
<depends on>—modifies a component depends on another component. For example, <Tomcat> <depends on> <Java>.
<connects to>—modifies a component that to connects or otherwise couples to another component. For example, <WebApp> <connects to> <Database>.
<is hosted on>—modifies a component may be hosted on another component. For example, <application> <is hosted on> <operating system>, <database> <is hosted on> <database management system>.
<is instance of> and <is subclass of> relationship types may be used to build hierarchies for reusability and maintenance. Properties defined at the higher level may be inherited by the lower level. A descendent node may have a <is a subclass of> relationship to its ancestor node.
<depends on>, <connects to> and <is hosted on> relationship types define the topology of the component during deployment. The object of the relationship may be deployed before the source of the relationship to ensure proper operation. The <is hosted on> relationship may imply a bonding between the source node and the target node. For example, the value of a property in the source node may be the same as that in the target node. The <connects to> relationship type may imply that the source node and the target node can run parallel when the nodes share another type of relationship, such as <depends on> or <is hosted on>.
Instance nodes may define dependent properties values and artifacts information to support proper operation. The flow generation logic may leverage the relationships captured in the layouts to automatically generate deployment workflows, such as a sequence of deployment scripts, that take into consideration of the requirements, dependencies, and policies used to support the deployment. The flow generation logic may accept a portion of the core layout and a one or more target nodes for deployment as inputs. The flow generation logic may produce one or more deployment workflows to implement the target nodes as an output.
In various implementations, the example pseudocode below may be used to implement a portion of flow generation logic on corresponding circuitry to support execution.
Referring back to
In the example user interface 2501, the first input 2510 is implemented as a combined drop-down menu and search bar. However, other input types, such as command lines, search bars, drop-down menus, button selectors, voice recognition, or other interface inputs, may be used to implement the first input 2510. The available service options may be defined by the services available in the service catalog. The user interface 2501 may further include inputs 2514, 2516, which may include selections dependent on the first input or independent of the first input.
In the example user interface 2501, appearance, available options, or other parameters for the inputs may be changed, e.g., by the configuration logic 300, as selections are made at the first input 2510. The changes may be defined, in part, by relationships and vertices in the core model layout. However, through service catalog selections or other onboarding processes the core model layout may be customized or otherwise altered. The user interface 2501 may further include a navigation bar 2504 for selection management functions including workflow development.
In the example user interface 2501, five options for services at the first input are shown. In the example, the services are related to platforms for managing content, logistics, client relationship data, and other data. The service options include StoreFront, SugarCRM, ProductCatalog, AccountService, and ContentRepository.
The definitions of the northbound and southbound services for the core layout may be the source for the initial structure and relationship information for the workflow layout. Referring again to
The initial structure and extensions may be shown in the workflow deployment display (WDD) 2630. The WDD 2630 may show a layout 2632 applicable to the workflow being developed in the user interface 2501. In the WDD 2630, the links between nodes may be seen. In some cases, relationships may be displayed. However, in the example WDD 2630 the relationships are not currently displayed. Some information may be omitted from the WDD 2630 to facilitate operator comprehension of the WDD 2630. The operator may access such information be interacting with the WDD (e.g., via human interface device (HID) selection actions, such as, clicking, touching, voice commands, gestures, or other HID selections actions). The WDD 2630 may be used for making selections for the workflow deployment within a layout view.
In the example structure shown in the WDD, SugarCRM 2682 is the current selection for the extension to the core model. StoreFront 2684 previously onboarded. Between SugarCRM 2682 and Apache 2686 and between SugarCRM 2682 and IIS 2688 is a “hosted on” relationship. Although not currently shown, this relationship may be viewed on the WDD 2630 via HID selection interaction with the links show between SugarCRM 2682 and Apache 2686 and between SugarCRM 2682 and IIS 2688. Apache 2686 and IIS 2688 are both available for selection.
In the example second frontend 2600, the operator has not yet provided selections for the “preference” input 2516. In the example user interface 2501, the “preference” input 2516 is marked “optional”. The “optional” marking may indicate to an operator that a workflow may be deployed without necessarily providing selection to that particular input. In some cases, “optional” markings may be added or removed response to selections made at previous inputs, e.g., the first input 2510 of the selections made in the WDD 2630 in this example scenario.
The current model may be saved using save input 2660. The selections may be cleared using clear input 2662. In some cases, the configuration logic 300 may access records, such as resource availability, operator history for the particular operator or other operators, cost (e.g., cloud services, software licensing, or other costs), regulations, security, service level agreements (SLAs), cloud provider agreements, facility location preferences, benchmark performance (e.g., latency, throughput, or other performance factors for middleware, firmware, software, or hardware), enterprise preferences, or other factors to determine likely workflow deployment selections. For example, likely choices may be determined by applying a default weighted policy enforcement scheme that incorporates operator selection history. The operator may interact with the “likely choice” input 2664 to select components and services based on these likely selections. Provision of such likely choice options may allow the operator to develop a workflow while providing an avenue to make viable selections, avoid duplication of previous work, or otherwise increase selection efficiency.
Certain portions of the core layout may be less likely to change than other portions. For example, base resource types for a given workflow deployment (e.g., resources specified by an IaaS provider) may be more stable than operator-specified resources, such as, application frameworks (e.g., middleware, language framework, or other frameworks), applications, and operating systems.
Moving to
In
The fourth example frontend 2900 may further include a layout model view 2920 that may show the core model layout 2921 and previous extensions 2922 with extensions for the currently developed workflow 2924. The fourth example frontend may also allow HID interaction with the running instances display 2502. The running instances display 2502 may be manipulated to show details on the currently developed workflow.
The user interface 2501 may also be used to select policies, e.g., applying cost constraints or optimizations, enforcing region preferences, applying performance metrics or optimizations, applying availability constraints or other policies, through the preferences input 2516.
In an example scenario a performance preference may be implemented. In the example scenario, an SQLServer hosted on Windows, which is hosted on AWSEC2, is selected. The SQLServer was selected over a competing MYSQL deployment because, in this case, the SQLServer deployment has better throughput. In the example scenario, the available options for the SQLServer deployment allow for a throughput of 10000 records/s, and the available options for MYSQL deployment allow for a throughput of 9000 records/s. Thus, in this case the SQLServer deployment was selected for the performance advantage. However, in other deployment scenarios different throughputs may be achievable with different platforms. For example, MYSQL may be selected over SQLServer for performance in other scenarios.
Policies may be given weights such that when there is a conflict between two policies the policy with higher weight may be given priority over the policy given less weight. For example if cost is given a higher weight than a concurrently enforce performance preference, the lowest cost resources may be selected. However, if two resources have the same cost, the higher performance resource may be selected. In some cases, weights may also be used to scale the relative gains from different policies-based selections. In an example scenario, a cost increase of 10% may be overridden by a performance increase of 5%. However, in the same example scenario, a cost increase of 20% is not necessarily overridden by a performance increase of 5%. Thus, different weights may be placed on different relative gains without necessarily causing selection in accord with one policy over another regardless of the size of their relative gains. The following pseudocode, which may be implemented as e.g., a SPARQL code, may be used to implement a middleware throughput policy:
Referring again to
In the example shown in
When the model is saved, a manifest representing the model may be generated. The manifest may be created in one of multiple different formats, such as scripting languages, markup languages, or other formats. For example, a manifest may be in generated in JSON, YAML, or other formats. In various implementations, the system may format the manifest according to platform constraints. Thus, when a model is reused the system may not necessarily generate a manifest in the same format as when the model was previously used. Thus, the system may select components of the layout to be included in the manifest and generate the manifest based on the characteristics of the layout. Hence, some systems generate the manifest from the layout and need not necessarily depend on translation of the manifest. However, some systems may translate a manifest from format to format.
In the manifest, the properties, deployment artifacts, interfaces, and implementation artifacts for the components of the layout (e.g., core layout and extensions, applicable deployment components, or other set of components) may be captured in the manifest. For example, properties may include minimum RAM (e.g., minRAM) for virtual machines, server root passwords for MYSQL. The relationships for the links between components may also be captured in the manifest, so that deployment may be orchestrated when the manifest is passed to the deployment engine.
The deployment engine may accept a manifest as in input. The deployment engine may setup a connection with the platform provider. The manifest may be implemented in parallel or sequentially with other manifests. However, where dependencies a present, dependent manifests may be implemented in sequence with the manifests on which they depend. In an example case, the deployment engine may be implemented via an AWS SDK and cloudFormation service. The deployment engine using the AWS SDK may accept JSON format manifests. The following pseudocode may be used to create AWS stack with CloudFormation JSON manifest.
In another example case, the deployment engine may be implemented using the Apache jClouds library. The jClouds library may be used with multiple cloud providers. The following pseudocode may be used by the Apache jClouds based deployment engine:
After the deployment engine deploys the instance of the layout in the manifest, the instance's properties may be written back into the layout. For example, virtual machine IP addresses may not necessarily be known prior to deployment. In some cases, e.g., AWS, data such a vpc-id and subnet-id may not necessarily be known a priori. The deployment engine may fetch such properties and update the layout. The following pseudocode, e.g., SPARQL code, may be used to write a fetched value IP address value (e.g., 10.1.1.5 in this example) back to the layout:
Some platform providers may use configuration management database (CMDB) to capture instance information. Hence, the system may fetch information using the existing CMDB functions. The fetched information may be then written back to the layout as described above.
In various implementations infrastructure layout architecture may be used to enforce policy compliance on a core layout and/or instances. For example, a given component of a northbound service may be fulfilled by a number of different southbound services. However, a policy may bar certain non-compliant southbound services from being used to fulfill the component. In an example scenario, for a connected car type northbound service, a messaging service may be a component. However, a policy, for example, may require that automotive connectivity services be secured. Therefore, to comply with the policy, the system may fulfill the messaging service using a messaging service that provides security, such as encryption.
The policy logic may determine if one or more policy models apply to the portion the layout (506). If no policies apply, the policy logic may return to activity monitoring (501). If one or more policies apply, the policy logic may determine a priority for the policies (508). If conflicting policies exist, the policy logic 500 may determine which policy to enforce (510). For example, business rules may be applied. In an example scenario, a newer policy may be applied over an older policy. Additionally or alternatively, security policies may be applied over performance policies or vice versa.
The policy logic 500 may determine to verify compliance with a selected policy (512). The policy logic may determine a source node and destination node for the selected policy (514). The policy logic may determine if matching nodes are present in the portion of the layout (516). If nodes matching the determine source and destination nodes are not present, the policy logic 500 may generalize definitions of nodes in the layout and preform the matching again. For example, a determined source node in a policy may specify a “vehicle”, but the vehicle type node found in the portion of the layout may specify a “boat”. The policy logic 500 may generalize the node in the layout to its type “vehicle” to match the node to that source node in the policy. In no nodes are found that can be generalized in the manner, the policy logic may return to monitoring (501) or determine another policy to verify (512) if multiple applicable policies were found. Once the source and destination nodes are matched, the policy logic may perform a source-to-destination length comparison (518). For example, the policy logic may determine if the same number of component and relationship “hops” are present between the source and destination nodes in the policy model and layout.
If the source-to-destination lengths match in the policy model and the layout from matched source node to matched destination node, the policy logic 500 may compare the intervening components and relationships in the policy model and the layout (520). If the components and relationships are the same, the policy logic 500 may indicate compliance with the policy model (522). For example, the policy logic 500 may take no further action in response to compliance. Additionally or alternatively, the policy logic 500 may send an indication of compliance to allow another action to go forward. For example, the policy logic 500 may send an indication of compliance the flow generation logic 800 to allow the deployment of an instance to go forward. In another example, the policy logic 500 may send an indication to the layout logic 700 to indicate that a given extension of a core layout may be implemented.
If the source source-to-destination lengths do not match in the policy model and the layout from matched source node to matched destination node, the policy logic may determine if expansions and/or generalizations may be applied to the intervening components in the layout (524). In some cases, a layout may supply a simple service and fulfillment path. For example, a messaging service may be fulfilled by a wireless carrier generating one hop in the layout. However, the one hop in the layout may imply one or more, lower hierarchy relationships and components. For example, the messaging service provided by the wireless carrier may imply many “hosted on” and/or security components. The policy logic 500 may expand these implied relationships and components (526) and determine a new length between the source and destination nodes (528). Additionally or alternatively, multiple hops may be generalized or simplified to one hop, where the multiple hops are implied. The policy logic 500 may generalize the expanded relationships and components (530) and determine a new length between the source and destination nodes (528). Once the source-to-destination lengths match (518), the policy logic may provide to component and relationship comparison (520). Additionally or alternatively, the policy logic 500 may apply expansions and generalizations to the policy model depending on the implementation. In various implementations, generalizations and/or expansions may be applied when the resultant components and relationships generate a matching pair between the layout and policy model. When a match is not created, the expansion and/or generalization need not necessarily be applied. If the lengths cannot be matched after available expansions and generalizations have been applied, the policy logic 500 may indicate non-compliance with the policy model (530).
In some cases, the policy logic 500 may monitor instances that are currently deployed and/or actively running. The policy logic 500 may determine if an instance is deployed and/or running (532). The policy logic 500 may compile a list of policies associated with the deployed and/or running instances (534). In some cases, a change in the core layout may occur while an instance is deployed. The policy logic 500, may evaluate policies on the compiled list for continued compliance (508-530).
If, after the comparison (520), the components and/or relationships do not match, the policy logic 500 may indicate non-compliance with the policy model (530). For example, a non-compliance alert may be generated. The policy logic 500 may halt the layout logic 700 when applying an extension to prevent the non-compliance. Alternatively, the policy logic may cause the layout logic 700 to apply an extension to fix the non-compliance. The policy logic 500 may send an indication that a possible adjustment to a layout may lead to non-compliance with one or more policies. In another example, the policy logic 500 may prevent deployment of an instance by the flow generation logic 800.
The flow generation logic 800 and the policy logic 500 may cooperate to determine compliance on one or more possible deployments with one or more policies. For example, the flow generation logic may determine one or more possible workflows for deployment and the policy logic may indicate policy compliance and/or non-compliance for the workflows. Thus, the user may select among deployments based on widest compliance and/or highest priority compliance.
Referring again to
In some implementations, the flow generation logic may receive an indication of a service for deployment (802). For example, the flow generation logic may receive a user selection or other compulsory service selection. The flow generation logic 800 may send the service indication to the policy logic 500 as a layout input (892). Responsive the layout inputs, the flow generation logic 800 may receive the policy compliance indications from the policy logic 500 (898). Based on the policy compliance indications, the flow generation logic 800 may generate a workflow (822), incorporating policy compliance and the external selection.
In some cases, an option to extend the core layout may be presented to the operator.
In various implementations, a selected option or newly applied extension may change the options available to the operator.
The memory 620, may be used to store the data and/or media for available layouts 662; extensions 663; policy models 664; business rules 665, service catalogs 666, northbound service definitions 667, and/or southbound service definitions 668 to support the configuration logic 300, layout logic 700, flow generation logic 800, and/or policy logic 500 described above.
In various implementations, the example execution environment 600 may connect to one or more service catalog databases 690 for access service catalogs for definitions and/or configuration details of various northbound and southbound services.
The execution environment 600 may also include commutation interfaces 612, which may support wireless, e.g. Bluetooth, Wi-Fi, WLAN, cellular (4G, LTE/A), and/or wired, ethernet, Gigabit ethernet, optical networking protocols. The communication interface may support communication with external or third-party servers and/or service catalog databases 690. The execution environment 600 may include power functions 634 and various input interfaces 628. The execution environment may also include a user interface 618 that may include human interface devices and/or graphical user interfaces (GUI). The GUI may be used to present a management dashboard, actionable insights and/or other information to the user. In various implementations, the GUI may support portable access, such as, via a web-based GUI. In various implementations, the system circuitry 614 may be distributed over multiple physical servers and/or be implemented as one or more virtual machines.
The layout logic 700 may be fully or partially contained in the querying layer 1030 below the business rules layer 1020. The querying layer 1030 may handle access, traversal and manipulation of various layouts, such as core layouts and extensions. In an example implementation, access, traversal and manipulation of the layout may be performed with layouts in a hierarchical graphical format, such as RDF.
The layout storage layer 1040 may handle storage of the layouts. The layout may be stored in the same format as that used for access, traversal and manipulation. However, other formats may be used for storage. For example, data may be stored in a non-graphical tabular triplet format and then transformed to a hierarchical and/or graphical format for access, traversal and manipulation.
The flow generation logic 800 and/or policy logic 500 may be fully or partially contained in the ontology layer 1050. The ontology 1050 layer may support application of policies and determination of deployment strategies. These operations may be performed in a manner opaque to the operator and the operator may be presented with options simplified through rules at the business rules layer 1020. In various implementations the operator may access detail at the ontology layer through options to assert or de-assert various business rules at the business rules layer 1020. Thus, the operator may view transparent operation of the ontology layer.
The deployment layer 1060 interfaces with the ontology, querying, and user interface layer to execute operations and/or workflows selected by the operator at the user interface layer 1010 and/or automatically selected by the upper layers 1030, 1050. The deployment layer 1060, may translate operations from upper layers 1010, 1030, 1050 into machine language instructions for execution.
The system may traverse the core data model 2020 to discover relationships. For instance, the individual user 2030 of the MMS System 2022 may inherit from users 2024 the concept of “Services Used.” The system may also traverse the model to obtain the Products 2026 that a user subscribes to and then the Services 2028 associated with that particular product. Note that the core data model 2020 contains available configuration options and their relationships. For instances, the core data model 2020 may define mandatory and optional Services 2028 associated with a Product 2026, or that a Product 2026 must include at least one Service 2028.
The composite model 2100 maps data and services across the MMS System 2020 model and the connected car model 2102. In the regard, the composite model also creates new entities, e.g., Automotive 2104 and Connect Car User 2106 to help connect data and services across models and, e.g., resolve mismatch in data and services between the models.
In this example, the home automation model 2402 includes a Connected Home entity 2406 that is the basis for Users 2408, Attributes 2410, and Services 2412. Users 2408 serves as a basis for the Owner entity 2414 and Utility entity 2416. Attributes 2410 serves as a basis for the Location entity 2418 and the Devices entity 2420, while Services 2412 serves as a basis for Connectivity 2422.
The infrastructure layout architecture may be used to support northbound services, such as, self-care portals, business support systems, application storefronts, payment gateways, application support (e.g. social media applications, catalogs, application mangers), mediation, converge subscription management, access support, transaction monitoring, network gateways, customer relationship management, and/or other northbound services.
The methods, devices, processing, and logic described above may be implemented in many different ways and in many different combinations of hardware and software. For example, all or parts of the implementations may be circuitry that includes an instruction processor, such as a Central Processing Unit (CPU), microcontroller, or a microprocessor; an Application Specific Integrated Circuit (ASIC), Programmable Logic Device (PLD), or Field Programmable Gate Array (FPGA); or circuitry that includes discrete logic or other circuit components, including analog circuit components, digital circuit components or both; or any combination thereof. The circuitry may include discrete interconnected hardware components and/or may be combined on a single integrated circuit die, distributed among multiple integrated circuit dies, or implemented in a Multiple Chip Module (MCM) of multiple integrated circuit dies in a common package, as examples.
The circuitry may further include or access instructions for execution by the circuitry. The instructions may be stored in a tangible storage medium that is other than a transitory signal, such as a flash memory, a Random Access Memory (RAM), a Read Only Memory (ROM), an Erasable Programmable Read Only Memory (EPROM); or on a magnetic or optical disc, such as a Compact Disc Read Only Memory (CDROM), Hard Disk Drive (HDD), or other magnetic or optical disk; or in or on another machine-readable medium. A product, such as a computer program product, may include a storage medium and instructions stored in or on the medium, and the instructions when executed by the circuitry in a device may cause the device to implement any of the processing described above or illustrated in the drawings.
The implementations may be distributed as circuitry among multiple system components, such as among multiple processors and memories, optionally including multiple distributed processing systems. Parameters, databases, and other data structures may be separately stored and managed, may be incorporated into a single memory or database, may be logically and physically organized in many different ways, and may be implemented in many different ways, including as data structures such as linked lists, hash tables, arrays, records, objects, or implicit storage mechanisms. Programs may be parts (e.g., subroutines) of a single program, separate programs, distributed across several memories and processors, or implemented in many different ways, such as in a library, such as a shared library (e.g., a Dynamic Link Library (DLL)). The DLL, for example, may store instructions that perform any of the processing described above or illustrated in the drawings, when executed by the circuitry.
Various implementations have been specifically described. However, many other implementations are also possible.
This application claims priority to provisional application Ser. No. 62/046,150, filed Sep. 4, 2014, which is entirely incorporated by reference.
| Number | Date | Country | |
|---|---|---|---|
| 62046150 | Sep 2014 | US |
