Petter Svärd

Evaluation of delta compression techniques for efficient live migration of large virtual machines

Despite the widespread support for live migration of Virtual Machines (VMs) in current hypervisors, these have significant shortcomings when it comes to migration of certain types of VMs. More specifically, with existing algorithms, there is a high risk of service interruption when migrating VMs with high workloads and/or over low-bandwidth networks. In these cases, VM memory pages are dirtied faster than they can be transferred over the network, which leads to extended migration downtime. In this contribution, we study the application of delta compression during the transfer of memory pages in order to increase migration throughput and thus reduce downtime. The delta compression live migration algorithm is implemented as a modification to the KVM hypervisor. Its performance is evaluated by migrating VMs running different type of workloads and the evaluation demonstrates a significant decrease in migration downtime in all test cases. In a benchmark scenario the downtime is reduced by a factor of 100. In another scenario a streaming video server is live migrated with no perceivable downtime to the clients while the picture is frozen for eight seconds using standard approaches. Finally, in an enterprise application scenario, the delta compression algorithm successfully live migrates a very large system that fails after migration using the standard algorithm. Finally, we discuss some general effects of delta compression on live migration and analyze when it is beneficial to use this technique.

High Performance Live Migration through Dynamic Page Transfer Reordering and Compression

Although supported by many contemporary Virtual Machine (VM) hyper visors, live migration is impossible for certain applications. When migrating CPU and/or memory intensive VMs two problems occur, extended migration downtime that may cause service interruption or even failure, and prolonged total migration time that is harmful for the overall system performance as significant network resources must be allocated to migration. These problems become more severe for migration over slower networks, such as long distance migration between clouds. We approach this two-fold problem through a combination of techniques. A novel algorithm that dynamically adapts the transfer order of VM memory pages during live migration reduces the risk of re-transfers for frequently dirtied pages. As the amount of transferred data is thereby reduced, the total migration time is shortened. By combining this technique with a compression scheme that increases the migration bandwidth the migration downtime is also reduced. An evaluation by means of synthetic migration benchmarks shows that our combined approach reduces migration downtime by a factor 10 to 20, shortens total migration time by around 35%, as well as consumes between 26% and 39% less network bandwidth. The feasibility of our approach for real-life applications is demonstrated by migrating a streaming video server 31% faster while transferring 51% less data.

Cost-Optimal Cloud Service Placement under Dynamic Pricing Schemes

Until now, most research on cloud service placement has focused on static pricing scenarios, where cloud providers offer fixed prices for their resources. However, with the recent trend of dynamic pricing of cloud resources, where the price of a compute resource can vary depending on the free capacity and load of the provider, new placement algorithms are needed. In this paper, we investigate service placement in dynamic pricing scenarios by evaluating a set of placement algorithms, tuned for dynamic pricing. The algorithms range from simple heuristics to combinatorial optimization solutions. The studied algorithms are evaluated by deploying a set of services across multiple providers. Finally, we analyse the strengths and weaknesses of the algorithms considered. The evaluation suggests that exhaustive search based approach is good at finding optimal solutions for service placement under dynamic pricing schemes, but the execution times are usually long. In contrast, greedy approaches perform surprisingly well with fast execution times and acceptable solutions, and thus can be a suitable compromise considering the tradeoffs between quality of solution and execution time.

Principles and Performance Characteristics of Algorithms for Live VM Migration

Since first demonstrated by Clark et al. in 2005, live migration of virtual machines has both become a standard feature of hypervisors and created an active field of research. However, the rich ongoing research in live migration focusmainly on performance improvements to well-known techniques, most of them being variations of the Clark approach. In order to advance live migration beyond incremental performance improvements, it is important to gain a deeper understanding of the live migration problem itself and its underlying principles. To address this issue, this contribution takes a step back and investigates the essential characteristics of live migration. The paper identifies five fundamental properties of live migration and uses these to investigate, categorize, and compare three approaches to live migration: precopy, postcopy and hybrid. The evaluated algorithms include well-known techniques derived from that of Clark as well as novel RDMA in-kernel approaches. Our analysis of the fundamental properties of the algorithms is validated by a set of experiments. In these, we migrate virtual machines with large memory sizes hosting workloads with high page dirtying rates to expose differences and limitations of the different approaches. Finally, we provide guidelines for which approach to use in different scenarios.

Self-management challenges for multi-cloud architectures

Addressing the management challenges for a multitude of distributed cloud architectures, we focus on the three complementary cloud management problems of predictive elasticity, admission control, and placement (or scheduling) of virtual machines. As these problems are intrinsically intertwined we also propose an approach to optimize the overall system behavior by policy-tuning for the tools handling each of them. Moreover, in order to facilitate the execution of some of the management decisions, we also propose new algorithms for live migration of virtual machines with very high workload and/or over low-bandwidth networks, using techniques such as caching, compression, and prioritization of memory pages.

A General Approach to Service Deployment in Cloud Environments

The cloud computing landscape has recently developed into a spectrum of cloud architectures, leading to a broad range of management tools for similar operations but specialized for certain deployment scenarios. This both hinders the efficient reuse of algorithmic innovations within cloud management operations and increases the heterogeneity between different management systems. Our overarching goal is to overcome these problems by developing tools general enough to support the full range of popular architectures. In this contribution, we analyze commonalities in recently proposed cloud models (private clouds, multi-clouds, bursted clouds, federated clouds, etc.), and demonstrate how a key management functionality - service deployment - can be uniformly performed in all of these by a carefully designed system. The design of our service deployment framework is validated through a demonstration of how it can be used to deploy services, perform bursting and brokering, as well as mediate a cloud federation in the context of the OPTIMIS Toolkit.

Continuous Datacenter Consolidation

Efficient mapping of Virtual Machines~(VMs) onto physical servers is a key problem for cloud infrastructure providers as hardware utilization directly impacts profit. Today, this mapping is commonly only performed when new VMs are created, but as VM workloads fluctuate and server availability varies, any initial mapping is bound to become suboptimal over time. We introduce a set of heuristic methods for continuous optimization of the VM-to-server mapping based on combinations of fundamental management actions, namely suspending and resuming physical machines, migrating VMs, and suspending and resuming VMs. By using these methods, cloud infrastructure providers can continuously optimize their server resources regardless of the predictability of the workload. To verify that our approach is applicable in real-world scenarios, we build a proof-of-concept datacenter management system that implements the proposed algorithms. The feasibility of our approach is evaluated through a combination of simulations and real experiments where our system provisions a workload of benchmark applications. Our results indicate that the proposed algorithms are feasible, that the combined management approach achieves the best results, and that the VM suspend and resume mechanism has the largest impact on provider profit.

REST-Based SOA Application in the Cloud: A Text Correction Service Case Study

In this paper, we present a REST-based system, \emph{Set It Right} (\emph{SIR}), where people can get feedback on and help with short texts correction. The rapid development of the SIR system, enabled by designing it as a set of services, and also leveraging commercially offered services, illustrates the strength of the SOA paradigm. Finally, we evaluate the cloud computing techniques and infrastructures used to deploy the system and how cloud technologies can help shorten the time to market and lower the initial costs.

High performance fault-tolerance for clouds

Cloud computing and virtualized infrastructures are currently the baseline environments for the provision of services in different application domains. While the number of service consumers increasingly grows, service providers aim at exploiting infrastructures that enable non-disruptive service provisioning, thus minimizing or even eliminating downtime. Nonetheless, to achieve the latter current approaches are either application-specific or cost inefficient, requiring the use of dedicated hardware. In this paper we present the reference architecture of a fault-tolerance scheme, which not only enhances cloud environments with the aforementioned capabilities but also achieves high-performance as required by mission critical every day applications. To realize the proposed approach, a new paradigm for memory and I/O externalization and consolidation is introduced, while current implementation references are also provided.

Hecatonchire: Towards Multi-host Virtual Machines by Server Disaggregation

Horizontal elasticity through scale-out is the current dogma for scaling cloud applications but requires a particular application architecture. Vertical elasticity is transparent to applications but less used as scale-up is limited by the size of a single physical server. In this paper, we propose a novel approach, server disaggregation, that aggregates memory, compute and I/O resources from multiple physical machines in resource pools. From these pools, virtual machines can be seamlessly provisioned with the right amount of resources for each application and more resources can be added to vertically scale a virtual machine as needed, regardless of the bound of any single physical machine. We present our proposed architecture and implement key functionality such as transparent memory scale-out and cloud management integration. Our approach is validated by a demonstration using benchmarks and a real-world big-data application and results indicate a low overhead in using memory scale-out in both test cases.