With the proliferation of cloud based services - IaaS, PaaS, and SaaS - enterprises are increasingly consuming all type of IT services delivered by cloud providers. These cloud services are increasingly being used to create integrated solutions where some of the component services are delivered from on-premise private cloud environments and the remaining are provided by off-premise providers. While this best-of-breed approach results in flexible and agile business solutions, it also raises significant problems related to metering, billing, charge back and accounting. In this paper, after reviewing common hybrid cloud integration patterns, we discuss the importance of usage metering and the associated challenges in hybrid cloud environments. We then present a novel solution for metering of services delivered from multiple cloud providers. In this approach, service metering is keyed off of the life-cycle events generated by the service management controls across the hybrid cloud. By identifying life-cycle events associated with services consumed, service usage is tracked from all sources and then filtered and aggregated in a centralized manner. The aggregated information can then be used in an on-line manner for policy-based service delivery, charge-back, billing and reporting, and auditing. Based on this approach we have developed an end-to-end service usage metering and billing system for hybrid cloud environments. Some of the key underlying technologies have been incorporated in IBM SmartCloud Orchestrator - an IBM offering for managing workload patterns in multi cloud environments.

Adoption of hybrid clouds by enterprises has been hampered by the inability of current hybrid cloud infrastructures to provide scalable and efficient mechanisms (1) to ensure the trustworthiness and integrity of the software stack executing a hybrid application workload, or (2) to enforce governmental privacy, data jurisdiction and audit regulations by ensuring that remote data and computation do not cross specified geographic boundaries.

From an enterprise perspective, one key motivation to transform the traditional IT management into Cloud is the cost reduction of the hosted services. In an Infrastructure-as-a-Service (IaaS) Cloud, virtual machine (VM) instances share the physical machines (PMs) in the provider's data center. With large number of PMs, providers can maintain low cost of service downtime at the expense of higher infrastructure and other operational costs (e.g., power consumption and cooling costs). Hence, determining the optimal PM capacity requirements that minimize the overall cost is of interest. In this paper, we show how a cost analysis and optimization framework can be developed using stochastic availability and performance models of an IaaS Cloud. Specifically, we study two cost minimization problems to address the capacity planning in an IaaS Cloud: (1) what is the optimal number of PMs that minimizes the total cost of ownership for a given downtime requirement set by service level agreements? and, (2) is it more economical to use cheaper but less reliable PMs or to use costlier but more reliable PMs for insuring the same availability characteristics? We use simulated annealing, a well-known stochastic search algorithm, to solve these optimization problems. Results from our analysis show that the optimal solutions are found within reasonable time.

In this paper, we describe architecture and design of an innovative integrated monitoring solution for monitoring workloads in a hybrid cloud. Using off-the-shelf commercially available monitoring products and services, the solution described here allows for monitoring of a workload deployed across public and private clouds separated by enterprise firewall. The monitored data from endpoints distributed across the hybrid cloud is aggregated in a centralized monitoring system for analysis and for performing higher level management functions. Our solution is automated and can be integrated with other IT service management functions. The solution can leverage any endpoint monitoring functions and services a cloud service provider may offer. The monitoring framework described in this paper forms the basis for the hybrid cloud monitoring capability in IBM's hybrid cloud offering,IBM Service Management Extensions for Hybrid Cloud.

One of the major security threats that public cloud computing platforms face today is that the active cloud virtual machine instances are visible and accessible via the public internet, which allows hackers to carry out several types of attacks such as Denial of Service (DoS) and intrusion over along durations which increases the probabilities of successful penetration. Security logs of the failed attempts attest to the real threat and the intensity and duration of these. Most systems running on public cloud instances today are not security hardened to withstand such persistent and long attacks. It is not only dangerous but also disastrous for the enterprise that uses such instances to deliver cloud services, for the users that use such services, and for the cloud provider that provides the cloud infrastructure. Therefore, what is required is a network level access control solution that facilitates delivery of cloud services while protecting the network perimeter of the solution in a useable and dynamically customisable manner. In this paper, we have described such a network-based access control solution for public cloud services that we have designed and developed and is applicable to any of the various cloud platforms available today. We have deployed our solution as part of the "Security-as-a-Service" model on IBM Smart Cloud Enterprise (SCE), and has been used for commercial delivery of cloud services. These applications have led to not only high level of security with no security attacks via network exposure on the services, but also significant savings on the cost of maintaining the security of such instances and services. We have also studied the challenges that network address translators (NATs) pose for network-based access control on public cloud, and have developed solutions for such challenges.

In this paper, we first describe the challenges and pain points of adopting off-premise cloud-based computing services by enterprise users. To address these challenges, we have developed a hybrid cloud architectural framework for controlling and managing network of integrated computing services in on- and off-premise cloud environments. We identify three types of integration patterns that are commonly observed and describe support in the hybrid cloud framework for such patterns. The framework allows creation, modification, and management of integrated hybrid cloud services. Using this framework, we describe how solutions can be designed for policy-based access to cloud services from on-premise environments and for policy-based secure access to on premise data from off-premise cloud based services. The framework offers capabilities for (i) on-demand capacity expansion or cloud-bursting, (ii) service composition and management across multiple cloud domains, (iii) unification and customization of service offerings from multiple cloud service providers, (iv) seamless integration of common workload management services such monitoring, metering, and security, and (v) unified governance of IT operation across the hybrid cloud. We then describe a realization of this architecture that has served as the basis of IBM's hybrid cloud solution offerings and describe how the hybrid cloud framework described here mitigates some of cloud adoption risks and lowers or eliminates the inhibitors.

Cloud service providers are constantly looking for ways to increase revenue and reduce costs either by reducing capacity requirements or by supporting more users without adding capacity. Over-commit of physical resources, without adding more capacity, is one such approach. Workloads that tend to be 'peaky' are especially attractive targets for over-commit since only occasionally such workloads use all the system resources that they are entitled to. Online identification of candidate workloads and quantification of risks are two key issues associated with over-committing resources. In this paper, to estimate the risks associated with over-commit, we describe a mechanism based on the statistical analysis of the aggregate resource usage behavior of a group of workloads. Using CPU usage data collected from an internal private Cloud, we show that our proposed approach is effective and practical.

High availability is one of the key characteristics of Infrastructure-as-a-Service (IaaS) cloud. In this paper, we show a scalable method for availability analysis of large scale IaaS cloud using analytic models. To reduce the complexity of analysis and the solution time, we use an interacting Markov chain based approach. The construction and the solution of the Markov chains is facilitated by the use of a high-level Petri net based paradigm known as stochastic reward net (SRN). Overall solution is composed by iteration over individual SRN sub-model solutions. Dependencies among the sub-models are resolved using fixed-point iteration, for which existence of a solution is proved. We compare the solution obtained from the interacting sub-models with a monolithic model and show that errors introduced by decomposition are insignificant. Additionally, we provide closed form solutions of the sub-models and show that our approach can handle very large size IaaS clouds.

Handling diverse client demands and managing unexpected failures without degrading performance are two key promises of a cloud delivered service. However, evaluation of a cloud service quality becomes difficult as the scale and complexity of a cloud system increases. In a cloud environment, service request from a user goes through a variety of provider specific processing steps from the instant it is submitted until the service is fully delivered. Measurement-based evaluation of cloud service quality is expensive especially if many configurations, workload scenarios, and management methods are to be analyzed. To overcome these difficulties, in this paper we propose a general analytic model based approach for an end-to-end perform ability analysis of a cloud service. We illustrate our approach using Infrastructure-as-a-Service (IaaS) cloud, where service availability and provisioning response delays are two key QoS metrics. A novelty of our approach is in reducing the complexity of analysis by dividing the overall model into sub-models and then obtaining the overall solution by iteration over individual sub-model solutions. In contrast to a single one-level monolithic model, our approach yields a high fidelity model that is tractable and scalable. Our approach and underlying models can be readily extended to other types of cloud services and are applicable to public, private and hybrid clouds.

Cloud based services may experience changes - internal, external, large, small - at any time. Predicting and quantifying the effects on the quality-of-service during and after a change are important in the resiliency assessment of a cloud based service. In this paper, we quantify the resiliency of infrastructure-as-a-service (IaaS) cloud when subject to changes in demand and available capacity. Using a stochastic reward net based model for provisioning and servicing requests in a IaaS cloud, we quantify the resiliency of IaaS cloud w.r.t. two key performance measures - job rejection rate and provisioning response delay.

We describe a hybrid linear programming (LP) and evolutionary algorithm (EA) based resource matcher suitable for heterogeneous grid environments. The hybrid matcher adopts the iterative approach of the EA methods to perform a goal oriented search over the solution space and, within each iteration, uses the LP method to solve a partial resource matching problem. By judiciously controlling the partial problem size and its complexity, the hybrid matcher balances the accuracy of the solution and the execution time. We describe a grid management architecture that incorporates the hybrid resource matcher. Performance results indicate that the execution time of the hybrid matcher, under a variety of conditions, is at least as good and often significantly better than the execution time of LP and EA based matchers. The hybrid matcher is found to scale well with the complexity of the problem and to maintain sensitivity to the response time constraints of on-line environments.

Resource visualization has emerged as a powerful technique for customized resource provisioning in grid and data center environments. In this paper, we describe efficient strategies for policy-based controlling of virtualization of the physical resources. With these strategies, visualization is controlled taking into account workload requirements, available capacities of physical resources, and the governing policies. Realizing this control requires simultaneous handling of three problems: (i) determining the virtual resource configurations, (ii) the mapping of resulting virtual resources to physical resources, and (iii) the mapping of workloads to the virtual resources. We pose this as an optimization problem and solve this problem using a linear programming (LP) based approach. We evaluate this approach by implementing it in the Harmony grid environment consisting of heterogeneous resources and heterogeneous workload. Experimental results indicate that our approach is efficient and effective. We extend this approach further by using a two-phase heuristic that allows the decision making component to scale up to handle large scale grid systems.

We describe a novel approach for the on-line scheduling of storage network access in shared application-hosting environments. Our storage resource management and reservation service (RMRS) dynamically makes advance reservation of storage bandwidth for various accesses while resolving the potential for conflict unutilized storage network capacity is maintained at an optimal level for being able to support unexpected load/contingencies, while preventing any capacity underutilization. Among the novelties offered by our work are: (l)a Web-service architecture (2) an application domain of storage accesses for normal database workload, backup, replication, migration and recovery, while supporting business objectives, QoS, and service-level commitments (3) a formulation of the scheduling problem as a practical linear programming (LP) problem comprising a reshaping linearization and integer relaxation of the modeled non linear bandwidth constraints.

In this paper, we first present a linear programming based approach for modeling and solving the resource matching problem in grid environments with heterogeneous resources. The resource matching problem described takes into account resource sharing, job priorities, dependencies on multiple resource types, and resource specific policies. We then propose Web service style architecture for online matching of independent jobs with resources in a grid environment and describe a prototype implementation. Our preliminary performance results indicate that the linear programming based approach for resource matching is efficient in speed and accuracy and can keep up with high job arrival rates of an important criterion for online resource matching systems. Also, the Web service style architecture makes the system scalable and extendable. It can also be integrated with other existing grid services in a straightforward manner.

In this paper, we describe a new scheme for checkpointing parallel applications on message-passing scalable distributed memory systems. The novelty of our scheme is that a checkpointed application can be restored, from its checkpointed state, in a reconfigured form. Thus, a parallel application may be checkpointed while executing with t1 tasks on p1 processors, and then restarted from the checkpointed state with t2 tasks on p2 processors. As a result, applications can recover from partial failures in the underlying system. Also, the reconfigurable checkpointed states can be migrated from one parallel system to another even if they do not have the same number of processors. We describe a new programming model for implementing a reconfigurable checkpointing scheme for parallel programs. This new model is derived from the DRMS programming model, developed in the context of run-time reconfiguration of parallel applications. A key component of our implementation is the distribution-independent representation of application array data structures in persistent storage. For further optimizing the performance of checkpoint/restart operations, we provide parallel array section streaming operations for such distributed arrays. We present performance data for the reconfigurable checkpointing and restarting of parallel applications and compare that with the performance of conventional forms of checkpointing. Our results demonstrate the advantages of the new scheme we describe.

In this paper, we describe efficient, scalable, and deadlock-free asynchronous communication strategies suitable for unstructured computations on iPSC/860. Using these deadlock-free strategies, which incur small overhead, we have optimized the communication in parallel sparse Cholesky factorization. We present experimental results to show that such optimizations can lead to more than 50% saving in communication costs.

We consider an important tradeoff between processor and memory allocation in distributed parallel processing systems. To study this tradeoff, we formulate stochastic models of parallel program behavior, distributed parallel processing environments and memory overheads incurred by parallel programs as a function of their processor allocation. A mathematical analysis of the models is developed, which includes the effects of contention for shared resources caused by paging activity. We conduct a detailed analysis of real large-scale scientific applications and use these results to parameterize our models. Our results show that memory overhead resulting from processor allocation decisions can have a significant effect on system performance in distributed parallel environments, strongly suggesting that memory considerations must be incorporated in the resource allocation policies for parallel systems. We also demonstrate the importance of the inter-locality miss ratio, which is introduced in this paper and analyzed for the first time.

The authors analyze three general classes of scheduling policies under a workload typical of large-scale scientific computing. These policies differ in the manner in which processors are partitioned among the jobs as well as the way in which jobs are prioritized for execution on the partitions. The results indicate that existing static schemes to not perform well under varying workloads. Adaptive policies tend to make better scheduling decisions, but their ability to adjust to workload changes is limited. Dynamic partitioning policies, on the other hand, yield the best performance and can be tuned to provide desired performance differences among jobs with varying resource demands.

The authors propose a Cholesky factorization scheme based on a static task and communication schedule generated by symbolic preprocessing of the intercolumn dependencies. This information is used to reduce the overheads of maintaining data structures and the communication costs incurred in message passing during the numerical factorization step. The authors introduce three primitives which encapsulate these optimizations and alleviate the user's programming effort. Performance results obtained on the iPSC/860 show that an implementation using these primitives results in 30% to 40% savings over one that does not use any static communication structure information.>

A block based, automatic partitioning and scheduling methodology is presented for sparse matrix factorization on distributed memory systems. Using experimental results, this technique is analyzed for communication and load imbalance overhead. To study the performance effects, these overheads were compared with those obtained from a straightforward 'wrap mapped' column assignment scheme. All experimental results were obtained using test sparse matrices from the Harwell-Boeing data set. The results show that there is a communication and load balance tradeoff. The block based method results in lower communication cost whereas the wrap mapped scheme gives better load balance.

For multiprocessor systems with two level memory hierarchy; the communication requirements of parallel Cholesky factorization of dense and sparse symmetric, positive definite matrices are analyzed. The data traffic associated with computing the Chloesky factor of an nn ×n dense matrix using nα processors, α ≤ 2, is shown to be Ω(n2+α/2), assuming that the computational load is uniformly distributed. For an nn ×n sparse matrix, representing a √n × √n regular grid graph, the corresponding data traffic is shown to be Ω(n1+α/2), α ≤ 1.