Patent Application
20130339203
The present disclosure relates to a methodology for risk-based dynamic geo-location based replication of services in cloud computing and a system for implementing the same.
Cloud computing provides storage, compute, and other information technology (IT) services on demand. Over the years, many organizations have either moved all, or part of, their applications and services to a cloud to provide, or employed cloud solutions to dynamically adjust, the IT infrastructure by integrating computing services according to surges and peak demands.
Dynamic provisioning of resources is employed to replicate capabilities and/or services in a distributed computing infrastructure to overcome potential disruptions in the capabilities and/or services. Predictive tools for weather forecasts, risk profile analysis based on geographical location of data/service centers, and historical data are employed to improve service resiliency. Further, for each local computing service that is considered for replication, the cost of disruption is compared with the total cost of replication to ensure that a computing infrastructure service provider is selected in a cost-efficient manner.
According to an aspect of the present disclosure, a method of dynamically provisioning resources for a distributed computing infrastructure is provided. The method includes evaluating a risk, under at least one predicted environmental event, of not replicating at least one local computing service that an operator of the distributed computing infrastructure is expected to provide. The method further includes determining whether to replicate the at least one local computing service by adding at least one computing infrastructure service provider (CISP) to the distributed computing infrastructure based on evaluation of the risk. In addition, the method further includes adding at least one selected CISP to the distributed computing infrastructure if a determination to add the selected CISP is made.
According to another aspect of the present disclosure, a system for dynamic provisioning of resources for a distributed computing infrastructure is provided. The system includes one or more processor units in communication with a memory, and is configured to perform a method. The method includes a step of evaluating a risk, under at least one predicted environmental event, of not replicating at least one local computing service that an operator of the distributed computing infrastructure is expected to provide. The method further includes a step of determining whether to replicate the at least one local computing service by adding at least one computing infrastructure service provider (CISP) to the distributed computing infrastructure based on the evaluation of the risk. The method yet further includes a step of adding at least one selected CISP to the distributed computing infrastructure if a determination to add the selected CISP is made.
According to another aspect of the present disclosure, a system for dynamic provisioning of resources for a distributed computing infrastructure is provided. The system includes a computational resource selection module including a risk analysis module configured to evaluate a risk, under at least one predicted environmental event, of not replicating at least one local computing service that an operator of the distributed computing infrastructure is expected to provide, and configured to determine whether to replicate the at least one local computing service by adding at least one computing infrastructure service provider (CISP) to the distributed computing infrastructure based on the evaluation of the risk. The system further includes a service replication and computational infrastructure allocation module configured to receive instructions from the computational resource selection module and to add at least one selected CISP to the distributed computing infrastructure if the computational resource selection module generates an instruction to add the at least one selected CISP.
According to yet another aspect of the present disclosure, a non-transitory machine readable data storage medium embodying a computer program for dynamically provisioning resources for a distributed computing infrastructure is provided. The computer program includes instructions for performing a step of evaluating a risk, under at least one predicted environmental event, of not replicating at least one local computing service that an operator of the distributed computing infrastructure is expected to provide. The computer program further includes instructions for determining whether to replicate the at least one local computing service by adding at least one computing infrastructure service provider (CISP) to the distributed computing infrastructure based on evaluation of the risk. In addition, the computer program includes instructions for adding at least one selected CISP to the distributed computing infrastructure if a determination to add the selected CISP is made.
As stated above, the present disclosure relates to a methodology for risk-based dynamic geo-location based replication of services in cloud computing and a system for implementing the same. Aspects of the present disclosure are now described in detail with accompanying figures. The drawings are not necessarily drawn to scale.
As used herein, “cloud computing” refers to the delivery of computing hardware, computing software, and/or storage capacity as a service to a heterogeneous community of end-recipients.
As used herein, a “cloud” refers to a set of all infrastructures employed to provide a service of cloud computing.
As used herein, “grid based computing” or “grid computing” refers to a form of distributed and parallel computing, whereby a virtual computer includes a cluster of networked, loosely coupled computers acting in concert to perform very large tasks.
As used herein, an “alternative infrastructure” refers to any infrastructure that is not part of a cloud with respect to which the alternative infrastructure is referred.
As used herein, a “computing service” can be any service that cloud computing can provide.
As used herein, a “local computing service” is a computing service provided within a geographically limited region smaller than the entire earth.
In broad terms, a system and a method to dynamically replicate services of a distributed computing system on an alternative infrastructure are provided according to embodiments of the present disclosure. As used herein, a distributed computing system refers to any system including multiple autonomous computers that communicate through a computer network in order to achieve a common goal. Distributed computing systems include, for example, cloud and grid based computing systems. Dynamical replication of services of a distributed computing system can compensate for, and/or mitigate the damage caused by, predicted disruptions of a company's physical computing infrastructure. Dynamical replication of services of the distributed computing system can use alternative computing infrastructure technologies, environmental forecasts, data centre geographic location information, and historical data, in order to increase service availability and delivery.
According to an aspect of the present disclosure, the dynamic provisioning capabilities that characterize distributed computing infrastructure such as cloud and grid computing services are leveraged to elastically shape the company's physical infrastructure and provide resiliency as a service.
A feature of the present disclosure employed to provide the dynamic provisioning capabilities includes risk-based analysis of service replication or migration.
Another feature of the present disclosure employed to provide the dynamic provisioning includes the use of environmental event forecasts, such as weather forecasts, and geographic location information in order to determiner where services need to be replicated and to select a computing infrastructure service provider. This feature can be useful in case of natural and man-made disasters, which are usually geographically located. Other environmental event forecasts can include, but are not limited to, natural phenomena, such as storms, bushfires, and floods. Such man-made or natural environmental events can create disruptions that might make system's services unavailable or unreachable. For example, interruption of communication lines and network connections can cause the system services unavailable. Because many of the environmental events are geographically located, it is possible to minimize the impact of the environmental events by making potentially affected services redundant on demand.
Yet another feature of the present disclosure employed to provide the dynamic provisioning capabilities includes the use of historical data on environmental events (such as weather events) and service demands to aid the evaluation of a potential site (or location) to which services must be replicated or migrated.
Referring to
The system can include a computational resource selection module (which is also referred to as a “cloud selection module” in embodiments applied to cloud computing). The computational resource selection module can include a service information management module, an availability prediction module, an infrastructure monitoring module, and a risk analysis module. Each of these modules can include one or more processors in communication with a memory and configured to run programming instructions for performing various steps that the module enables.
Further, the system includes a service replication and computational infrastructure allocation module, which is configured to receive instructions from the computational resource selection module and to add at least one selected computing infrastructure service provider (CISP) to the distributed computing infrastructure. In one embodiment, the distributed computing infrastructure can be provided as a cloud computing infrastructure configured to provide one of more of infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). As used herein, a CISP refers to any service provider that provides a computing infrastructure service in the form or Iaas, PaaS, SaaS, or any combination thereof. A CISP may be within a same company as the operator of the distributed computing infrastructure, or can be a different company from the operator of the distributed computing infrastructure.
Infrastructure as a service is a cloud service model in which a cloud provider offers computers (as physical machines or as virtual machines), storage, firewalls, load balancers, and networks. Platform as a service is a cloud service model in which a cloud provider delivers a computing platform and/or solution stack typically including operating system, programming language execution environment, database, and web server. Software as a service (SaaS) is a cloud service model in which a cloud provider installs and operates application software in the cloud and cloud users access the software from cloud clients. The various services that a cloud provides can be accessed by various computing devices known in the art, including, but not limited to, servers, notebooks, desktops, tablet PC's, and phones.
The system of
The computation resource selection module can select a location or infrastructure service in which services are to be replicated employing the various modules therein. Referring to operation 10 of
In one embodiment, the distributed computing infrastructure is a cloud computing infrastructure. In one embodiment, the services provided by the operator of the distributed computing infrastructure through at least one computing infrastructure service provider (CISP) can include at least one local computing service. As used herein, a “local computing service” is a service that an operator is obligated to provide within a geographically limited region that is smaller than the entire world. The geographically limited region may be within a continent, within a country, within a state, or within any geographically defined region. In one embodiment, the at least one local computing service includes at least one of cloud application service, cloud platform service, and cloud infrastructure service.
Referring to operation 20 of
Because each of the local computing services is limited within the corresponding geographically limited region, each predicted environmental event is analyzed to determine the corresponding region of impact, i.e., areas that are affected by the predicted environmental event. In one embodiment, the availability prediction module can perform calculations only when the region of impact affects a local computing service. The availability prediction module can calculate predicted values of availability given the input that specifies various parameters of the predicted environmental effect.
Environmental event forecasts are provided as the input to the availability prediction module. Further, the availability prediction module can include, or be in communication with, a database on historical data on environmental events. The database on historical data on environmental events can include, for example, historical information on weather conditions, seasonal characteristics, and past disruptions. Employing the environmental event forecasts and the database on historical data on environmental events, the availability prediction module can provide forecasts on the effects on the distributed computing infrastructure and the local computing services of the predicted environmental events within the region of impact of the predicted environmental events. For example, the availability prediction module can make availability predictions for each candidate computational infrastructure (such as cloud sites or data centers) and the local computing services within the region of impact of the predicted environmental events. For example, an availability index can be derived based on previous experience on the impact of certain weather events on information technology (IT) infrastructure of the considered region. For instance, strong seasonal raining that affects specific geographic regions can be considered to determine the availability index.
In one embodiment, the availability prediction module can be configured to invoke the operation of the risk analysis module only if an estimated probability of disruption of the at least one local computing service is greater than a predefined value.
Referring to operation 30 of
Referring to operation 40 of
For example, under predicted extreme weather conditions, the computational resource selection module can invoke the risk analysis module to evaluate the risk of replicating or not replicating a given service. The risk of replicating includes the risk of incurring excessive cost during the process of replicating. The risk of not replicating includes the risk of a service disruption and accompanying financial and non-financial losses to the operator of the distributed computing environment.
Referring to operation 50 of
In one embodiment, the risk analysis module is configured to determine an impact of disruption of the at least one local computing service to the operator, and to determine an estimated total cost of replicating the at least one local computing service by adding the at least one computing infrastructure service provider. Further, the risk analysis module can be configured to determine the probability at which the predicted environmental event is expected to disrupt the at least one local computing service. For example, risk assessment techniques, often used for IT security purposes, in which the incurred risk depends on threats, vulnerability and asset/service value can be adapted to compute the risk of replicating (or not replicating) a service set to an alternate infrastructure. In such a case, inclement weather conditions can be factored in as a potential threat that can compromise the well functioning of the IT infrastructure in consideration.
In one embodiment, the risk analysis module can be configured to compute an expectation value for a total financial cost of not replicating the local computing service, and to compare the expectation value with the estimated total cost of replicating the at least one local computing service. Any of the methods described above can be employed to compute the expectation value for the total financial cost of not replicating the local computing service.
In one embodiment, the risk analysis module can be configured to include, within the estimated total cost of replicating the at least one local computing service, the cost of adding the at least one computing infrastructure service provider to the distributed computing infrastructure, and the cost of operating the at least one computing infrastructure service provider for a duration of the at least one predicted environmental event.
In one embodiment, the risk analysis module can be configured to generate a list of available alternate computing infrastructure service providers (CISP's) for each selected local computing service among the at least one local computing service, and to compute a replication time needed to replicate the selected local computing service for each available alternate CISP among the list of available alternate CISP's. Further, the risk analysis module can be configured to calculate the estimated time until initiation of a possible disruption due to the at least one predicted environmental event. In addition, the risk analysis module can be configured to determine an estimated total cost of replicating the at least one local computing service only for available alternate CISP's having a replication time that is less than the estimated time until initiation of the possible disruption.
In one embodiment, the risk analysis module can be configured to compute an estimated duration of a possible disruption from at least one predicted environmental event. Further, the risk analysis module can be configured to determine an estimated total cost of replicating the at least one local computing service based on the computed estimated duration of the possible disruption.
In one embodiment, the risk analysis module can be configured, for each selected local computing service among the at least one local computing service, to generate a list of available alternate CISP's, and to compute a minimum total cost of replicating the selected local computing service employing an alternate CISP among the list of available alternate CISP's. The list of available alternate CISP's can be selected based on the geo-location of the CISP's so that the services provided by the CISP's within the list are not affected by the predicted environmental event.
In one embodiment, the risk analysis module can be configured to compute the minimum total cost by calculating, for each available alternate CISP among the list of available alternate CISP's, a total cost of replicating the selected local computing service employing the each available alternate CISP, and by selecting a minimum value among the calculated total costs of replicating the selected local computing service.
In one embodiment, the risk analysis module can be configured to include, within the total cost of replicating the selected local computing service to an available alternate CISP under consideration, the cost for replicating the selected local computing service to the available alternate CISP under consideration, and the cost for restoring the selected local computing service from the available alternate CISP under consideration.
In one embodiment, the risk analysis module can be configured, for each selected local computing service among the at least one local computing service, to generate an indexed list of available alternate CISP's based on predetermined specified criteria or predetermined business constraints. Further, the risk analysis module can be configured to set an initial value for an index for the indexed list of available alternate CISP's at an extremum (e.g., a minimum or a maximum), and to increment (in case the index is initially set at the minimum) or to decrement (in case the index is initially set at the maximum) the index until an available alternate CISP is found that is capable of providing the selected local computing service at a total cost of replicating the selected local computing service that is less than an expectation value for a total financial cost of not replicating the local computing service, or until all CISP's within the indexed list of available alternate CISP's are examined.
In one embodiment, the selection and replication processes consider the location of candidate CISP's, the amount of time that is required to move data and code from a cloud infrastructure to be replicated to a selected alternative infrastructure, the estimated loss in case of disruption, and the cost incurred in using the provider's services for the planned amount of time.
In one embodiment, the computational resource selection module can employ an algorithm that continually monitors the arrival of information on predicted environmental events such as weather forecasts. Once a forecast for a predicted environmental event is received, the availability prediction module can compute the likelihood of this event causing a disruption. If this likelihood is above a predefined threshold value, which can be specified by the system administrator or derived by the system based on historical data, the risk analysis module can compute the time left until the disruption and the duration of the disruption.
Subsequently, for each service and infrastructure service providers, the risk analysis module computes the cost of disruption, replication and restore. The risk analysis module can perform an analysis to check whether a replication to each infrastructure service provider is advantageous in various aspects including timing, reliability, robustness, and cost.
Referring to
Referring to step 410, an indexed list of local computing services that is potentially affected by a predicted environmental event can be generated. For example, flooding is a condition that has affected many densely populated areas over the past few years, especially in developing countries. According to an embodiment of the present disclosure, upon a forecast of such an event, a list of infrastructures that are likely to be compromised can be determined, and a list of services that currently depend on such infrastructure for operation can be determined. The services in the list can be the candidates for replication.
Referring to step 412, an estimated time until initiation of a possible disruption from the forecast for each local computing service can be computed based on the forecast on the nature of the predicted environmental event. For example, the estimated time until initiation of a possible disruption can be calculated employing the data generated by the availability prediction module based on the parameters of the predicted environmental events (e.g., beginning of a severe weather condition or arrival of a tsunami).
Referring to step 414, the estimated duration of the possible disruption can be computed from the forecast for each local computing service. For example, the estimated duration of the possible disruption can be calculated employing the data generated by the availability prediction module based on the parameters of the predicted environmental events.
Referring to step 416, a list of available alternate computing infrastructure service providers (CISP) can be generated, for example, from publicly available database (e.g., phone book directory), a database (not shown) configured to store information on alternate CISP's, and/or by manual entry of information. The list of available CISP's can be indexed employing any algorithm known in the art.
Referring to step 418, the index for the local computing services can be set to a minimum value. Alternately, any systematic index changing method can be employed provided that all of the local computing services affected by the predicted environmental event can be processed during the subsequent steps to provide adequate service replication as determined by comparison of the total cost of replication and the cost of not replicating, i.e., the value of disruption of each service.
Referring to step 425, a determination can be made as to whether replication for the selected local computing service has already been determined. If the decision on whether to replicate the local computing service corresponding to the current index value has already been made, the process flow proceeds to step 426.
At step 426, a determination is made as to whether the index for the local computing services is at the maximum. If the indexing scheme employs any other algorithm than incrementing the value of the index from the minimum, a determination can be made as to whether there exists any local computing service for which a determination on whether to replicate the local computing service has not yet been made. If the value of the index is at the maximum, or alternately, if there is no other local computing service for which a decision on whether to replicate the local computing service has not yet been made, the process flow proceeds to step 499, at which the process flow terminates.
If the value of the index is not at the maximum, the process flow proceeds to step 428, at which the value of the index for the local computing services is incremented to the next value. Alternately, a new local computing service for which a decision on whether to replicate the local computing service has not yet been made is selected if the indexing scheme employs any other algorithm than incrementing the value of the index from the minimum.
If the decision on whether to replicate the local computing service corresponding to the current index value has not already been made, the process flow proceeds from step 425 to step 430. At step 430, an estimated cost of disruption to the selected local computing service for the expected duration of disruption can be computed. For example, such cost of disruption can include, but not limited to, the financial loss caused by service unavailability and the damage of reputation incurred by the organization offering the service and the provider of the IT infrastructure hosting the service.
Referring to step 432, an indexed list of CISP's can be generated based on specified criteria or business constraints such as cost for unit of resource, maximum resource capacity and infrastructure utilization. In general, before iterating the list of hosted services and candidate computational infrastructures (i.e., alternate CISP's), the list of available alternate CISP's can be sorted according to a set of criteria specified by the system administrator, or based on predefined programmed business constraints. For example, the list of available alternate CISP's can be sorted by decreasing order of availability, increasing order of cost or increasing likelihood of being affected by the weather event in consideration. The CISP index can be set to a minimum at step 432.
Referring to step 440, the replication time needed to replicate the selected local computing service employing the selected CISP can be calculated for the selected CISP, i.e., for the CISP corresponding to the current value for the CISP index.
Referring to step 445, the replication time calculated for the selected local computing service employing the selected CISP (the CISP corresponding to the current value of the CISP index) is compared with the expected time of initiation of disruption of the selected local computing service. A determination can be made as to whether there is enough time for replication, i.e., whether the replication time calculated for the selected local computing service employing the selected CISP is less than the expected time of initiation of disruption of the selected local computing service.
If there is not enough time for replication, the process flow proceeds to step 456, at which a determination is made as to whether the CISP index is at the maximum value. If the CISP index is not at the maximum value, i.e., if the CISP index can be incremented, the process flow proceeds to step 448, at which the CISP index is incremented to the next value. The process flow then proceeds to step 440 with the incremented value for the CISP index.
If the CISP index is at the maximum value at step 456, the process flow proceeds to step 459, at which a determination is made not to replicate the selected local computing service. The process flow then proceeds to step 466, at which a decision is made as to whether the index for the local computing services is at the maximum. Alternately, if the indexing scheme employs any other algorithm than incrementing the value of the index from the minimum, a determination can be made as to whether there exists any local computing service for which a determination on whether to replicate the local computing service has not yet been made.
If at step 466, it is determined that the index for the local computing services is at the maximum (or that decision on whether to replicate the local computing services have been made for all local computing services under consideration), the process flow then proceeds to step 499, at which the process flow terminates. If step 466 determines that the index for the local computing services is not at the maximum (or that there exists at least one local computing service for which a decision on whether to replicate needs to be made), the process flow proceeds to step 468, at which the index for the local computing service is incremented to the next value. The process flow then proceeds to step 425.
If a determination is made that there is enough time for replication at step 445, the process flow proceeds to step 450, at which the cost of replicating the selected local computing service to the selected CISP is computed.
Referring to step 452, the cost of restoring the local computing service from the selected CISP can be computed.
The total cost of replicating the selected local computing service to the selected CISP (which is one of the available alternate CISP's) under consideration includes the cost for replicating the selected local computing service to the selected CISP under consideration, and the cost for restoring the selected local computing service from the selected CISP under consideration.
Referring to step 455, a comparison is made between the total cost of replicating the selected local computing service to the selected CISP and the cost of disruption to the operator of the distributed computing infrastructure, i.e., the cost of disruption to the provider of the selected local computing service. It is noted that the operator of the distributed computing infrastructure may, or may not, be within the same company as the client of the selected local computing service.
If step 455 determines that the total cost of replicating the selected local computing service to the selected CISP is greater than the cost of disruption to the operator of the distributed computing infrastructure, the process flow proceeds to step 456. At step 456, a determination is made as to whether the CISP index is at the maximum value as described above, and the process flow proceeds to step 448 or to step 459 depending on whether the CISP index is at the maximum value.
If step 455 determines that the total cost of replicating the selected local computing service to the selected CISP is less than the cost of disruption to the operator of the distributed computing infrastructure, the process flow proceeds to step 460.
At step 460, a decision is made to replicate the selected local computing service to the selected CISP. The process flow then proceeds to step 466, at which a determination is made on whether the index for the local computing services is at the maximum (or that decision on whether to replicate the local computing services have been made for all local computing services under consideration) as described above. Depending on whether the index for the local computing services is at the maximum, the process flow proceeds to step 468 or to step 499.
Referring to operation 60 of
If criteria for replicating the services are met, the risk analysis module can cause the service replication and computational infrastructure allocation module to perform a corresponding replication operation. Specifically, the information on what services need to be replicated where, as determined by the cloud selection module, can be passed on to the service replication and computational infrastructure allocation module. For example, the decisions made at step 460 and step 466 of
Referring to operation 70 of
Referring to
The computational resource selection module of embodiments of the present disclosure can invoke a risk analysis module to determine whether to replicate the affected local computing services employing available alternate CISP's, i.e., available CISP's that are not part of the set of cloud infrastructures within the cloud. The available alternate CISP's are analyzed by the risk analysis module to determine a back-up CISP 10C into which a service is to be replicated, and to screen out unselected CISP's, which are not used to replicate services of the cloud. The back-up CISP 10C becomes part of an expanded cloud, which is represented by the set of double sold lines and the dotted line that encloses the back-up CISP 10C. In other words, the back-up CISP 10C is added as an additional cloud infrastructure to the cloud on a temporary basis until the probability of disruption of service due to the predicted environmental event is extinguished.
Referring to operation 80 of
Referring to operation 90 of
The system including the computational resource selection module and the service replication computational infrastructure allocation module can be implemented independently of the physical sites for the infrastructures of the distributed computing infrastructure. Thus, the system including the computational resource selection module and the service replication computational infrastructure allocation module can be either at the site that requires the replication service or at any other location that contains all the required data to trigger the replication.
The following scenario is provided as an illustration of application of the method of an embodiment of the present disclosure. If a heavy storm is forecast for the next two days and the estimated impact area encloses one of a company's data center, the system of embodiments of the present disclosure can be employed to select the most appropriate data centre where the service provided by the company's affected data center, and to prevent disruption of the services provided by the company's data center.
While the disclosure has been described in terms of specific embodiments, it is evident in view of the foregoing description that numerous alternatives, modifications and variations will be apparent to those skilled in the art. Various embodiments of the present disclosure can be employed either alone or in combination with any other embodiment, unless expressly stated otherwise or otherwise clearly incompatible among one another. Accordingly, the disclosure is intended to encompass all such alternatives, modifications and variations which fall within the scope and spirit of the disclosure and the following claims.
