Hypervisor-based fault tolerance (HBFT), a checkpoint-recovery mechanism, is an emerging approach to sustaining mission-critical applications. Based on virtualization technology, HBFT provides an economic and transparent solution. However, the advantages currently come at the cost of substantial overhead during failure-free, especially for memory intensive applications. This paper presents an in-depth examination of HBFT and options to improve its performance. Based on the behavior of memory accesses among checkpointing epochs, we introduce two optimizations, read fault reduction and write fault prediction, for the memory tracking mechanism. These two optimizations improve the mechanism by 31.1% and 21.4% respectively for some application. Then, we present softwaresuperpage which efficiently maps large memory regions between virtual machines (VM). By the above optimizations, HBFT is improved by a factor of 1.4 to 2.2 and it achieves a performance which is about 60% of that of the native VM.

P2P storage systems use replication to provide a certain level of availability. While the system must generate new replicas to replace replicas lost to permanent failures, it can save significant replication cost by not replicating following transient failures. However, in real systems, it is impossible to reliably distinguish permanent and transients failures, resulting in a tradeoff between high recovery cost and low data availability. In this paper, we analyze the use of timeouts as a mechanism to navigate this tradeoff. We address the challenging problem of how to choose a timeout to walk the fine line between causing unnecessary replication due to detection inaccuracy, and reducing availability due to detection delay. We conduct simulations based both on synthetic and real traces, and show that the performance of our selected timeout closely approximates the optimal performance that can be achieved by timeouts, and even that of an "oracle" failure detector.

The commonly agreed Zipf-like access pattern of Web work-loads is mainly based on Internet measurements when text-based content dominated the Web traffic. However, with dra-matic increase of media traffic on the Internet, the inconsis-tency between the access patterns of media objects and the Zipf model has been observed in a number of studies. An insightful understanding of media access patterns is essential to guide Internet system design and management, including resource provisioning and performance optimizations. In this paper, we have studied a large variety of media work-loads collected from both client and server sides in different media systems with different delivery methods. Through ex-tensive analysis and modeling, we find: (1) the object refer-ence ranks of all these workloads follow the stretched expo-nential (SE) distribution despite their different media systems and delivery methods; (2) one parameter of this distribution well characterizes the media file sizes, the other well char-acterizes the aging of media accesses; (3) some biased mea-surements may lead to Zipf-like observations on media access patterns; and (4) the deviation of media access pattern from the Zipf model in these workloads increases along with the workload duration. We have further analyzed the effectiveness of media caching with a mathematical model. Compared with Web caching under the Zipf model, media caching under the SE model is far less effective unless the cache size is enormously large. This indicates that many previous studies based on a Zipf-like assumption have potentially overestimated the media caching benefit, while an effective media caching system must be able to scale its storage size to accommodate the increase of media content over a long time. Our study provides an analytical basis for applying a P2P model rather than a client-server model to build large scale Internet media delivery systems.

Instant messaging (IM) has become increasingly popular due to its quick response time, its ease of use, and possibility of multitasking. It is estimated that there are several millions of instant messaging users who use IM for various purposes: simple requests and responses, scheduling face to face meetings, or just to check the availability of colleagues and friends. Despite its popularity and user base, little has been done to characterize IM traffic. One reason might be its relatively small traffic volume, although this is changing as more users start using video or voice chats and file attachments. Moreover, all major instant messaging systems route text messages through central servers. While this facilitates firewall traversal and gives instant messaging companies more control, it creates a potential bottleneck at the instant messaging servers. This is especially so for large instant messaging operators with tens of millions of users and during flash crowd events. Another reason for the lack of previous studies is the difficulty in getting access to instant messaging traces due to privacy concerns. In this paper, we analyze the traffic of two popular instantmessaging systems, AOL InstantMessenger (AIM) and MSN/Windows Live Messenger, from thousands of employees in a large enterprise. We found that most instant messaging traffic is due to presence, hints, or other extraneous traffic. Chat messages constitute only a small percentage of the total IM traffic. This means, during overload, IM servers can protect the instantaneous nature of the communication by dropping extraneous traffic. We also found that the social network of IM users does not follow a power law distribution. It can be characterized by a Weibull distribution. Our analysis sheds light on instant messaging system design and optimization and provides a scientific basis for instant messaging workload generation.

This paper studies the conditions under which peer-to-peer (P2P) technology may be beneficial in providing IPTV services over typical network architectures. It has two major contributions. First, we contrast two network models used to study the performance of such a system: a commonly used logical "Internet as a cloud" model and a “physical” model that reflects the characteristics of the underlying network. Specifically, we show that the cloud model overlooks important architectural aspects of the network and may drastically overstate the benefits of P2P technology by a factor of 3 or more. Second, we provide a cost-benefit analysis of P2P video content delivery focusing on the profit trade-offs for different pricing/incentive models rather than purely on capacity maximization. In particular, we find that under high volume of video demand, a P2P built-in incentive model performs better than any other model for both high-definition and standard-definition media, while the usage-based model generally generates more profits when the request rate is low. The flat-reward model generally falls in-between the usagebased model and the built-in model in terms of profitability.

This paper presents a performance study of BitTorrent-like P2P systems by modeling, based on extensive measurements and trace analysis. Existing studies on BitTorrent systems are single-torrent based and usually assume the process of request arrivals to a torrent is Poisson-like. However, in reality, most BitTorrent peers participate in multiple torrents and file popularity changes over time. Our study of representative BitTorrent traffic provides insights into the evolution of single-torrent systems and several new findings regarding the limitations of BitTorrent systems: (1) Due to the exponentially decreasing peer arrival rate in a torrent, the service availability of the corresponding file becomes poor quickly, and eventually it is hard to locate and download this file. (2) Client performance in the BitTorrent-like system is unstable, and fluctuates significantly with the changes of the number of online peers. (3) Existing systems could provide unfair services to peers, where a peer with a higher downloading speed tends to download more and upload less. Motivated by the analysis and modeling results, we have further proposed a graph based model to study interactions among multiple torrents. Our model quantitatively demonstrates that inter-torrent collaboration is much more effective than stimulating seeds to serve longer for addressing the service unavailability in BitTorrent systems. An architecture for inter-torrent collaboration under an exchange based instant incentive mechanism is also discussed and evaluated by simulations.

This paper studies the conditions under which P2P sharing can increase the capacity of IPTV services over FTTN networks. For a typical FTTN network, our study shows a) P2P sharing is not bene cial when the total traffic in a local video office is low; b) P2P sharing increases the load on FTTN switches and routers in local video of ces; c) P2P sharing is the most beneficial when the network bottleneck is experienced in the southbound segment of a local video office (equivalently a northbound segment of an FTTN switch); and d) sharing among all FTTN serving communities is not needed when network congestion problems are solved by using some other technologies such as program pre-caching or replication. Based on the analytical results, we design and implement the MediaGrid platform for IPTV services which monitors FTTN network conditions and decides when and how to share videos among peers to maximize the service capacity. Simulations and bounds both validate the potential bene ts of the MediaGrid IPTV service platform.

Modern Internet streaming services have utilized various techniques to improve the quality of streaming media deliv-ery. Despite the characterization of media access patterns and user behaviors in many measurement studies, few stud-ies have focused on the streaming techniques themselves, particularly on the quality of streaming experiences they of-fer end users and on the resources of the media systems that they consume. In order to gain insights into cur-rent streaming services and thus provide guidance on de-signing resource-efficient and high quality streaming media systems, we have collected a large streaming media work-load from thousands of broadband home users and business users hosted by a major ISP, and analyzed the most com-monly used streaming techniques such as automatic protocol switch, Fast Streaming, MBR encoding and rate adaptation. Our measurement and analysis results show that with these techniques, current streaming systems tend to over-utilize CPU and bandwidth resources to provide better services to end users, which may not be a desirable and effective way to improve the quality of streaming media delivery. Moti-vated by these results, we propose and evaluate a coordina-tion mechanism that effectively takes advantage of both Fast Streaming and rate adaptation to better utilize the server and Internet resources for streaming quality improvement.

Streaming media objects have become widely used on the Internet, and the demand of interactive requests to these objects has increased dramatically. Typical interactive requests include fast forward and direct jumps. Unfortunately, most of existing streaming proxies are designed for sequential accesses, and only a few solutions have been proposed to maintain additional data structures in the proxy to support some interactive operations (such as fast forward) other than jumps, which are among the most common interactive requests from the clients. Focusing on interactive accesses, in this paper we present an analysis of streaming media workload collected from thousands of broadband users hosted by a major ISP. Our analysis shows that jump accesses (48%) and pauses (51%) are the dominant client interactive requests and that jump accesses often suffer serious delays due to slow buffering through the network. To support jump accesses effectively, we further propose a novel caching algorithm-DISC (Dynamic Interleaved Segment Caching), which trades cache performance for response time to client interactive requests. In this algorithm, segments of a media object are cached dynamically according to client access patterns. DISC can support direct jumps efficiently while ensuring timely prefetching of uncached segments for sequential accesses. Trace-driven simulations demonstrate that DISC outperforms other caching schemes significantly for interactive requests with only a small degradation in cache performance.

In this paper, we study the media workload collected from a large number of commercial Web sites hosted by a major ISP and that collected from a large group of home users connected to the Internet via a well-known cable company. Some of our key ndings are: (1) Surprisingly, the majority of media contents are still delivered via downloading from Web servers. (2) A substantial percentage of media downloading connections are aborted before completion due to the long waiting time. (3) A hybrid approach, pseudo streaming, is used by clients to imitate real streaming. (4) The mismatch between the downloading rate and the client playback speed in pseudo streaming is common, which either causes frequent playback delays to the clients, or unnecessary traffic to the Internet. (5) Compared with streaming, downloading and pseudo streaming are neither bandwid the cient nor performance effective. To address this problem, we propose the design of AutoStream, an innovative system that can provide additional previewing and streaming services automatically for media objects hosted on standard Web sites in server farms at the client's will.