Patent Application
20110145393
As 2008 comes to an end, the software computing world is shifting with the introduction of Cloud computing where Amazon (EC2/S3/SimpleDB), Microsoft (Azure/Services platform) and others are offering fully managed software execution resources such as CPU, storage, network, message queues, database etc. as commodity resources to be purchased in various models such as pay per use and monthly fee. These Cloud resources are targeted to fit low end requirements as well as scale to serve high end requirements of full scale internet applications.
This opens up an obvious opportunity for startups to bootstrap their operation with low investment while being able to scale their application as their business grows.
On the other end this shift also opens up an opportunity for organizations of all sizes from the small business to the enterprise to optimize their software computing resources while increasing scale, availability, agility and other important aspects such as disaster recovery.
For example: say that a large company would like to launch a new service for consumers. The company starts off with provisioning a few servers in their data center for this service. No knowing whether the service would be very successful, the company does not want to invest capital in purchasing a large amount of computers and bandwidth in advance. Rather, the company can use cloud computing resources if the service becomes very popular.
Another example is where a company need a large amount of software execution resources on a regular cycle of seasonal peak time or for a specific one time calculation. These are cases where the availability of a scale of software execution resources for a given period of time can both allow the company to deliver on its business as well as save considerable cost.
Last, consider the opportunity for organizations to move from dealing with physical computer hardware in their server farms to using software execution resources for running some or all of their business's software. This allows organizations to have flexible and scalable yet affordable software execution resources while delivering to business needs such as manageability, security, compliance, availability and infrastructure resiliency to disasters
A prudent organization might tap into software execution resources from multiple Cloud computing providers as well as have its own server farm to optimize on total cost of ownership and deliver on critical business functionality.
One of the conclusions from the above is that software execution resources across Cloud and on premises, will become a commodity that can be reserved, sold and utilized through various business models where both Cloud computing resources and on premises resources play into the equations as well as other parameters such as scale, reliability, availability etc.
When regarding software execution resources as a commodity, we can further view that this is a highly dynamic commodity since software execution is dependent on a myriad of factors and it can exceed the expected resources, be delayed, migrated to execute on other resources or even cancelled. It is also a commodity with high versatility of specification ranging from CPU, memory, storage to security, compliance and availability.
As a commodity, the software execution resources can be reserved through contracts and re-sold or re-negotiated as the business parameters for these resources change
The use of the software execution resources commodity can also be monitored to provide dynamic re-assessment of the software execution resources so that resource reservation contracts can be dynamically re-negotiated and both the provider of the resources and the consumer of the resources can monitor the use of the commodity
Thus, there is a need for a method for dynamic reservation of cloud and on premises resources for software execution based on receiving the specification for the resources needed for the software, evaluating the availability of resources and reserving the resources for the software, followed by monitoring the usage of the software execution resources.
A method for dynamic reservation of cloud and on premises resources for software execution may include a protocol to discover, evaluate availability, negotiate terms and create a reservation contract for software execution resources based on the required specification
A method for dynamic reservation of cloud and on premises resources for software execution may further include resource agent modules that represent available software execution resources (e.g.: CPU, memory, storage, network . . . ) and reservation agent modules that are used for communicating with the resource agent modules for reserving software execution resources. In addition, monitoring agent modules are used to monitor the actual software execution reservation contract.
A method for dynamic reservation of cloud and on premises resources for software execution may be used to implement a system for dynamic optimized reservation of software execution resources across cloud and on premises resources based on excess and demand and on online monitoring and re-negotiation of the software execution resources usage and reservation contracts
This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
In
Software executing in this environment can be utilizing a combination of resources from one or more of the cloud providers 101, 102 and on premises computing center 103 resources. For example—software can use the CPU/memory resources on premises 103a along with storage from Cloud computing provider 1's storage 101b with networking resources from both on premises 103c and Cloud computing provider 101c for communication between the two. This is a valid scenario for software that requires highly available distributed shared storage across multiple on premises computing centers.
The resource agents 201, 202, 203 are used to represent the available computing resources and negotiate the reservation of the computing resources of the computing environments 101, 102 and 103 respectively.
In this implementation, each of the resource agents may represent all or a subset of the computing resources in its environment and has the ability to process a request to reserve computing resources, determine whether resources are available, negotiate the terms for the reservation and authorize the reservation of the resources for a specific or unlimited timeframe. In addition, the resource agent has the ability to limit the entities that have the ability to reserve computing resources. For example, in this system, resource agent 103 accepts only on-premises requests.
An exemplary request to reserve computing resources, may include: The timeframe for the reservation, range of hardware specification, scale specification for the various resources such as CPU, Network, Storage . . . , the availability, security, price, pricing model, indemnification, cancellation fee etc. . . .
The reservation agents 211, 212, 213 are used for reserving computing resources. These reservation agents communicate with the various resource agents (e.g.: resource agent 201, resource agent 202 . . . ) and negotiate resource reservation according to resource allocation requirements.
Reservation agent 212 is a reservation agent with additional capabilities to re-negotiate reservations dynamically. This means that reservation agent 212 will make a reservation and continue to query the resource agents either on a schedule basis and/or when specific event occur in order to re-negotiate improved terms which could be price, availability and other services. When re-negotiating, the reservation agent will take into account the cost of “breaking” a reservation that was already made (since it is switching to a new reservation) and if the software is already executing on the current resources, the cost of switching (if any) to the new resources.
Reservation agent 213 is an on premises reservation agent which has the ability to reserve on-premises computing resources as well as cloud based computing resources. In this implementation, reservation agent 213 is not different than reservation agent 211, it just has access to additional resource agents (e.g.: resource agent 203). Reservation agent 213 is also a re-negotiating agent, thus as an example, in this implementation—when ample on premises computing resources become available, it might re-negotiate a reservation so that software running on a cloud computing environment can be moved to run on the available on premises resources.
In this implementation, the resource agents registers with a resource agent discovery web service 231 so that reservation agents can discover registered resource agents. As is obvious, there might be multiple web services to discover resource agents and these can be further federated. In addition, a reservation agent can have additional ways to discover resource agents such as DNS queries, multicast queries or a simple list provided to the reservation agent.
In this implementation, the resource contract aggregation agent includes three different distinct capabilities. The first capability is the ability to receive and maintain contracts of resources that are available to be reserved. Once a reservation is made for the available resources represented in the contracts that it maintains, the resource contract aggregation reports the reservation to the appropriate resource agent for the allocation of the actual computing resources.
The second capability is to present a resource agent interface to reservation agents so that reservation agents can reserve the available computing resources that are available in the contracts maintained by the resource contract aggregation agent. It can be observed that a resource contract aggregation agent can use different combinations of the computing resources available in the contracts it received to fulfill the request of the reservation agents
The third capability is to act as a reservation agent when a request for reserving computing resources is being made and there is no adequate available resource contract. The resource contract aggregation agent may then federate the request to another resource contract aggregation agent or to a resource agent to try and fill the requested reservation.
In addition, the monitoring information that is populated to web service 252 can be used by reservation agents when making resource reservation so that these reservation agents can take into account the reliability of the execution of the resource reservation contracts by the various computing environments
Last, in this implementation, the monitoring information populated to web service 252 can be used in real-time by re-negotiating reservation agents in a way that if a reservation contract is not being properly fulfilled, the re-negotiation reservation agent can reserve additional/other computing resources so that the appropriate computing resources will be immediately available for the executing software to meet its business requirements
Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.
