Dingding Li

A New Disk I/O Model of Virtualized Cloud Environment

In a traditional virtualized cloud environment, using asynchronous I/O in the guest file system and synchronous I/O in the host file system to handle an asynchronous user disk write exhibits several drawbacks, such as performance disturbance among different guests and consistency maintenance across guest failures. To improve these issues, this paper introduces a novel disk I/O model for virtualized cloud system called HypeGear, where the guest file system uses synchronous operations to deal with the guest write request and the host file system performs asynchronous operations to write the data to the hard disk. A prototype system is implemented on the Xen hypervisor and our experimental results verify that this new model has many advantages over the conventional asynchronous-synchronous model. We also evaluate the overhead of asynchronous I/O at host, which is brought by our new model. The result demonstrates that it enforces little cost on host layer.

A Performance Optimization Mechanism for SSD in Virtualized Environment

Applications in the cloud computing era have the emergent requirement of fast I/O support. Compared with a hard drive disk, a solid state disk (SSD) has low delay, low energy consumption, high throughput and other advantages. However, the semantics of an SSD cannot be recognized by current virtual machine monitors. The trim instruction, which plays an important role in space management of an SSD, cannot be passed to the underlying SSD device in a virtualized environment. So, how to bridge the semantics gap between the application layer and virtualization layer for SSD devices should be an important problem. We propose Vtrim to solve the above problem in this paper. Vtrim monitors the operations in the virtual machine and sends the SSD semantics to the Domain 0 immediately. And then the semantics are translated into the operations of Domain 0, which can trigger the SSD’s local instructions. To improve the write performance in multiple guest operating systems, we set a Vtrim cache to buffer all instructions from the guests and flush them into an SSD in a well-scheduled way. Experiment results in the para-virtualization environments with Vtrim show that the random write performance is improved by even up to 100% and the average response time is reduced by up to 40%.

Improving disk I/O performance in a virtualized system

Desktop virtualization is a general solution for providing users with various working environments on a single physical machine. It is typically based on the virtual machine (VM) technology, which can provide smart sharing policies on the scarce hardware. Compared with native computing environments, however, VM may sacrifice the performance, especially in the I/O subsystem. Although certain methods have been proposed to improve the performance of networking and graphic processing, the performance of disk I/O is almost ignored. In this paper, we propose two methods to reduce the extra overhead on Xen hypervisor at different layers of its disk protocol stack. The experimental results show that on the average 21% of unnecessary CPU cycles can be saved from the Xen hypervisor, and when a high performance disk device is applied in the system, our proposed optimization techniques can improve the overall disk I/O performance by 15.6%-36.6% for all benchmarks used in the experiments. Finally, by evaluating our methods in a practical system of desktop virtualization, user-applications can achieve an improvement of 15.9%-23.7% over the original Xen hypervisor.

Improving write amplification in a virtualized and multimedia SSD system

Due to offering fast random-access disk I/O, it appears that solid-state drives (SSD), which is based on NAND flash memory, can suit well with the environment of cloud computing, especially for the cloud providing video streaming services. However, by investigating a practical virtual desktop system, where runs video streaming workloads, we find that importing this kind of SSDs into a virtualized system is not as simple as merely a mechanical replacement. Because a large proportion of disk I/O being included in the video streaming workload is write I/O, the inherent weaknesses of NAND flash memory, write amplification (WA), will be magnified in a guest operating system (OS). Worse, some useful remedies in a native OS become disabled or inefficient due to the interposition of hypervisor layer. This paper describes and analyzes these problems based on a practical virtual desktop system, and then proposes a tailor-made method to relieve them. By evaluating realistic user workloads and several typical benchmarks, the result shows that our method can effectively improve these problems in our virtualized SSD system.

A Network-Friendly Disk I/O Optimization Framework in a Virtualized Cloud System

In a typical virtualized cloud system, network-attached storage (NAS) is the common technology for storing the disk image files of virtual machine (VM). Although the client/server model on NAS is more reliable and manageable, a concurrent accessing behavior from multiple VMs on the shared NAS server may exhaust the underlying network bandwidth, thus hurting the disk I/O performance of VMs. In this paper, we improve this problem by a set of optimizations, which are designed around bridging the semantic gap between the file system of VM and NAS server. The experimental result shows that our work can effectively reduce the network traffic between the NAS server and client. Meanwhile, the disk I/O performance of VMs can be also improved.

FIOS: a flexible virtualized I/O subsystem to alleviate interference among virtual machines

Serving as the infrastructure of cloud computing, virtualization technologies have attracted considerable interest in recent years for their excellent resource utility, scalability, and high availability. Virtual Machine Monitor (VMM), which is a key element in cloud computing, enables multiple guest operating systems running simultaneously to share the same physical resources. This may lead to significant interference of disk I/O performance among virtual machines (VM). Particularly, the I/O performance of none I/O intensive domains can be seriously injured by the advent of I/O intensive nodes. We address this problem by building a block-level cache in the virtualized layer to absorb I/O requests from different domains. This method not only effectively alleviates the I/O performance interference caused by I/O intensive domains, but also greatly improves the I/O performance of guest OS. We implement and evaluate a Flexible I/O Subsystem (FIOS) within Xen VMM and show an evident reduction of I/O performance interference among virtual machines as well as a remarkable improvement of disk throughput.

A novel disk I/O scheduling framework of virtualized storage system

Modern data centers usually use virtual machine technology to host various big data applications in a single physical machine, not only enhancing the server utilization, but also providing them with the hardware-level isolation. However, in a typical virtualized environment an extra software layer called virtual machine monitor (VMM) is often interposed between the hardware resource and guest operating system (virtual machine, VM), shielding the specific user-process semantic inside a running VM. As a result, it obstructs the disk I/O scheduler of VMM to acquire the accurate information of a user-process (often a big data application), and thus proposes a challenge on the I/O request scheduling as well as the disk resource management at the granularity of VM user-process. Eventually, the disk I/O performance of a virtualized system is sub-optimal. This paper introduces an improved disk I/O scheduling framework for the virtualized system. It aims at bridging the semantic gap between physical disk I/O scheduler and VM user-process, providing a fair sharing of disk I/O resource among concurrent VMs. At the same time, it improves the overall disk I/O performance through a novel method for creating the image file of VM. Besides, an extra scheduling algorithm is proposed to further refine the storage performance. Finally, we implement these improvements on Xen hypervisor and conduct extensive experiments to verify our framework. The experimental result shows that our work improve the performance of read-intensive, write-intensive and mixed workloads up to 9, 10.7 and 20% respectively.

Triple-L: Improving CPS Disk I/O Performance in a Virtualized NAS Environment

Network-attached storage (NAS) provides cyberphysical systems (CPS) with the scalable, efficient, and reliable backing storage, such as the mobile virtual desktop based on cloud infrastructure. Within this storage architecture, virtual machine (VM) instances running in the NAS client usually receive data from the complex physical world and then persist them in the neat cyberspace in the NAS server. In this paper, we propose Triple-L to improve VM disk I/O performance in the NAS architecture. According to the specific storage semantic, Triple-L decouples the VM image file into several subfiles at the host layer and then selectively moves them into the NAS clients. In such a way, a VM disk I/O request may be proceeded locally in the NAS client, instead of walking the external networking path repetitively between NAS server and client. We have implemented Triple-L in a Xen-based NAS system. An accessory solution for dealing with storage failure and VM live migration on Triple-L is also discussed and evaluated. The experimental result shows that our work can effectively improve the disk I/O performance of VMs. Meanwhile, it brings moderate overhead for VM live migration.

Writeback throttling in a virtualized system with SCM

Storage class memory (SCM) has the potential to revolutionize the memory landscape by its non-volatile and byte-addressable properties. However, there is little published work about exploring its usage for modern virtualized cloud infrastructure. We propose SCM-vWrite, a novel architecture designed around SCM, to ease the performance interference of virtualized storage subsystem. Through a case study on a typical virtualized cloud system, we first describe why current writeback manners are not suitable for a virtualized environment, then design and implement SCM-vWrite to improve this problem. We also use typical benchmarks and realistic workloads to evaluate its performance. Compared with the traditional method on a conventional architecture, the experimental result shows that SCM-vWrite can coordinate the writeback flows more effectively among multiple co-located guest operating systems, achieving a better disk I/O performance without any loss of reliability.

VDB: Virtualizing the On-Board Disk Write Cache

In the cloud environment, virtual machine monitor (VMM) enables multiple virtual machines (VM) running simultaneously to share the same physical resources, such as disk and its on-board disk write-cache. However, for the reasons of high reliability, some virtual machines need to adopt write-through policy for disk write-cache, which is a key component to improve systems I/O performance. This causes serious write performance degradation for other data-intensive virtual machines which prefer a write-back cache. To satisfy the requirement of reliability and performance for different VMs, we adopt a disk write-through cache for the virtual machines with high reliability requirement and create a replaced write cache at the VMM layer for each virtual machine with high write performance demand. We also design and implement a prototype system called VDB within Xen. Our experimental results show that our system solves the coexistence problem of reliable virtual machines and I/O-bound virtual machines for the use of disk write-cache on the same virtualization platform. It can guarantee the reliability requirement and improve the write performance for different VMs.