Erez Hadad

Guaranteeing High Availability Goals for Virtual Machine Placement

The placement of virtual machines (VMs) on a cluster of hosts under multiple constraints, including administrative (security, regulations) resource-oriented (capacity, energy), and QoS-oriented (performance) is a highly complex task. We define a new high-availability property for a VM, when a VM is marked as k-resilient, as long as there are up to k host failures, it should be guaranteed that it can be relocated to a non-failed host without relocating other VMs. Together with Hardware Predictive Failure Analysis and live migration, which enable VMs to be evacuated from a host before it fails, this property allows the continuous running of VMs on the cluster despite host failures. The complexity of the constraints associated with k-resiliency, which are naturally expressed by Second Order logic statements, prevented their integration into the placement computation until now. We present a novel algorithm which enables this integration by transforming the k-resiliency constraints to rules consumable by a generic Constraint Programming engine, prove that it guarantees the required resiliency and describe the implementation. We provide some preliminary results and compare our high availability support with naive solutions.

VM Placement Strategies for Cloud Scenarios

The problem of Virtual Machine (VM) placement in a compute cloud infrastructure is well-studied in the literature. However, the majority of the existing works ignore the dynamic nature of the incoming stream of VM deployment requests that continuously arrive to the cloud provider infrastructure. In this paper we provide a practical model of cloud placement management under a stream of requests and present a novel technique called Backward Speculative Placement (BSP) that projects the past demand behavior of a VM to a candidate target host. We exploit the BSP technique in two algorithms, first for handling the stream of deployment requests, second in a periodic optimization, to handle the dynamic aspects of the demands. We show the benefits of our BSP technique by comparing the results on a simulation period with a strategy of choosing an optimal placement at each time instant, produced by a generic MIP solver.

FTS: a high-performance CORBA fault-tolerance service

This paper presents a lightweight CORBA fault-tolerance service called FTS. The service is based on standard portable features of CORBA, and in that respect is fully CORBA compliant, but does not follow the FT-CORBA specifications in areas where the authors felt the latter interfered with their other design goals. The service features a unique architecture, based on a new type of an object adaptor, called Group Object Adaptor (GOA). The service is portable, interoperable, and aims for simplicity and high-performance request processing. Moreover, the service supports network partitions, some aspects of non-deterministic processing, and mixing ORBs of different vendors in the same fault-tolerance infrastructure. The paper also presents an analysis of the differences between the service design and FT-CORBA, with the hope of stimulating a discussion about future improvements to the FT-CORBA standard.

Client-side enhancements using portable interceptors

Interceptors are a useful technique for extending the basic functionality provided by an Object Request Broker without changing its implementation, nor affecting client code. The recently proposed CORBA standard includes a new definition of portable interceptors. This definition is very powerful, yet not trivial to use. The authors review this definition, and discuss how it can be applied to provide client side enhancement for caching, load-balancing, flow control (quality of service), and fault-tolerant soft real-time. A recommendation to the OMG for a very minor change in the standard that could greatly simplify its usability is also provided.

Adaptive Batching for Replicated Servers

This paper presents two novel generic adaptive batching schemes for replicated servers. Both schemes are oblivious to the underlying communication protocols. Our novel schemes adapt their batching levels automatically and immediately according to the current communication load. This is done without any explicit monitoring or calibration of the system. Additionally, the paper includes a detailed performance evaluation

Symphony: Managing Virtual Servers in the Global Village

A virtual server is a server whose location in an internet is virtual; it may move from one physical site to another, and it may span a dynamically changing number of physical sites. In particular, during periods of high load, it may grow to new machines, while in other times it may shrink into a single host, and may even allow other virtual servers to run on the same host. This paper describes the design and architecture of Symphony, a management infrastructure for executing virtual servers in internet settings. This design is based on combining CORBA technology with group communication capabilities, for added reliability and fault tolerance.

Distributed wisdom: analyzing distributed-system performance - latency vs. throughput

In analyzing distributed-system performance, its not enough to include a latency breakdown (even though its the easiest performance measure to get). Instead, we should indicate next to each item whether the activity is CPU-bound or I/O-bound. Quite often, CPU-bound activities take a small portion of the overall latency but have a much bigger impact - throughput and indirectly on the overall latency.

A virtual resource placement service

Server, storage, and network virtualization and the growing adoption of cloud computing has expanded both the complexity and the value of intelligent allocation and management of data center resources. Resource allocation in a cloud environment is of fundamental importance. There are many competing goals, with differing priorities, that contribute to optimizing virtual resource allocation and placement including performance, reliability, security, energy, etc. We have developed an open extensible architecture to provide placement recommendations which allows for different independently developed Domain Managers to provide input/advice on placement. We have further developed the means to orchestrate the placement, ensuring that the required configuration actions be enacted both before and after the migration of the virtual machine. This paper explores the topic of providing a core placement calculation and orchestration architecture to facilitate management of workload demands in a cloud environment. We describe this architecture for placement services and orchestration, and present some results from a prototype implementation.

Using selective acknowledgements to reduce the memory footprint of replicated services

This paper proposes the use of Selective Acknowledgements (SACK) from clients to services as a method for reducing the memory footprint of replicated services The paper discusses the general concept of SACK in replicated services and presents a specific implementation of SACK for an existing replication infrastructure Performance measurements exhibiting the effectiveness of SACK are also presented.