Vicente Santonja

Analysis of self-similarity in I/O workload using structural modeling

Demonstrates that disk-level I/O requests are self-similar in nature. We show evidence (both visual and mathematical) that I/O accesses are consistent with self-similarity. For this analysis, we have used two sets of disk activity traces collected from various systems over different periods of time. In addition to studying the aggregated I/O workload that is directed to the storage system, we perform a structural modeling of the workload in order to understand the underlying causes that produce the observed self-similarity. This structural modeling shows that self-similar behavior can be explained by combining two different approaches: the on/off source model and Coxs model. The former applies to those processes that remain active during the whole trace, while the latter applies to sources that show a very short activity time.

Modeling and simulation of storage area networks

Storage area networks (SANs) are an emerging data communications platform which interconnects servers and storage devices (such as disks, disk arrays, and tape drives) to create a pool of storage that users can access directly. This networking approach reports benefits such as computer clustering, topological flexibility, fault tolerance, high availability, and remote management. In order to evaluate the performance of these systems it is necessary to have the adequate tools. Usually, performance evaluation may be based on analytical modeling or simulation. Each of them differs in their scope and applicability. However the simulation modeling technique offers more freedom, flexibility, and accuracy than the analytical methods. Thus, when evaluating the performance of SANs, simulation modeling should be used. In this paper the issues involved in the modeling and design of a very flexible and easy to use SAN simulator are presented. This tool is able to consider among others, both real-world I/O traces and synthetic I/O traffic, message packetization, faults in links and switches, virtual channels, different routing algorithms, etc. We describe its main internal organization, the basic modeling mechanisms the simulator is based on, the main input parameters and output performance variables. Also, the analysis of preliminary results using I/O traces is presented, showing that the storage network increases self-similarity of the traffic received by servers, latency variations are more important for control messages than for data messages, and links have a low utilization.

Self-similarity in I/O workload: analysis and modeling

Recently, the notion of self-similarity has been applied to wide-area and local-area network traffic. This paper demonstrates that disk-level I/O requests are self-similar in nature. We show evidence, both visual and mathematical, that the I/O accesses are consistent with self-similarity. Moreover, we show that this property of I/O accesses is mainly due to writes. For our experiments, we use two sets of traces that collect the disk activity from two systems over a period of two months. Such behavior has serious implications for the performance evaluation of storage subsystem designs and implementations, since commonly-used simplifying assumptions about workload characteristics (e.g. Poisson arrivals) are shown to be incorrect. Using the ON/OFF model, we implement a disk request generator. The inputs of this generator are the measured properties of the available trace data. We analyze the synthesized workload and confirm that it exhibits the correct self-similar behavior.

A new approach in the modeling and generation of synthetic disk workload

Shows a new approach to generate synthetic disk workload. The work presented is based on previous results obtained from the analysis of real disk traces. The proposed disk workload generation model can capture the heavy-tailed behavior of real disk workload, a critical feature to reproduce disk subsystem congestion. The generator provides synthetic workload much more accurately than commonly-used models. Since the workload plays a critical role in performance evaluations, having a more accurate disk workload generator is important for storage researchers in order to obtain fair and unbiased performance predictions.

Dynamic power saving in fat-tree interconnection networks using on/off links

Current trends in high-performance parallel computers show that fat-tree interconnection networks are one of the most popular topologies. The particular characteristics of this topology, that provide multiple alternative paths for each source/destination pair, make it an excellent candidate for applying power consumption reduction techniques. Such techniques are being increasingly applied in computer systems and the interconnection network is not an exception, since its contribution to the system power budget is not negligible. In this paper, we present a mechanism that dynamically switches on and off network links as a function of traffic. The mechanism is designed to guarantee network connectivity, according to the underlying routing algorithm. In this way, the default routing algorithm can be used regardless of the power saving actions taken, thus simplifying router design. Our simulation results show that significant network power consumption reductions can be obtained at no cost. Latency remains the same although the number of operating network links is dynamically adjusted.

Power saving in regular interconnection networks built with high-degree switches

Nowadays, high-degree switches are available as building blocks of the interconnection network of clusters of PCs. An alternative to take advantage of the high number of switch ports is to connect every pair of switches through not only one but also several links (this is known as link trunking in other environments). This extra connectivity can be exploited by using adaptive routing algorithms, thus improving network throughput and reducing network congestion. However with low traffic loads, all the links that compose the trunk link will not be utilized, but this idle links continue consuming power. Power consumption reduction techniques are being applied everywhere in computer systems and the interconnection network is not an exception, as its contribution is not negligible. In this paper, we present a mechanism that dynamically switches on and off network links as a function of traffic. It is specially targeted to those networks where trunk links are used. The mechanism can switch off any link, provided that network connectivity is guaranteed, (i.e. every pair of switches should be connected through at least one active link). Indeed, this restriction makes possible to use the same routing algorithm regardless the power saving actions taken, thus simplifying router design. Our simulation results show that the network power consumption can be greatly reduced, at the expense of some increase in latency. Nevertheless, it is shown that the power reduction is always higher that this latency increases.

On the switch architecture for fibre channel storage area networks

The fast growth of data intensive applications has caused a change in the traditional storage model. The server-to-disk approach is being replaced by storage area networks (SANs), which enable storage to be externalized from servers, thus allowing storage devices to be shared among multiple servers. Nowadays, the majority of SANs use fibre channel. The standard for fibre channel defines several issues related to the switch interface, but does not make any suggestion about the internal switch architecture to be implemented by manufacturers. We analyze the key architectural switch characteristics for building fibre channel storage area networks. To do so, our starting point is the performance analysis of two different switch architectures, identifying their strongest and weakest points, and thus taking advantage of the best features from both of them. After this first analysis, we introduce several other features in the switch, concluding with a proposed architecture that doubles network throughput while reducing response delay.

On the effect of link failures in fibre channel storage area networks

The fast growth of data intensive applications has caused a change in the traditional storage model. The server-to-disk approach is being replaced by storage area networks (SANs), which enable storage to be externalized from servers, thus allowing storage devices to be shared among multiple servers. The prominent technology for implementing SANs is Fibre Channel, due to its suitability for storage networking. Although the probability of a link failure for individual links in a SAN is very low, this probability dramatically increases as the network size becomes larger. Moreover, there are external factors, such as accidental link disconnections, that also can affect the overall SAN reliability. Until the faulty element is replaced, the SAN is functioning in a degraded mode. In this paper we analyze by simulation the performance degradation of Fibre Channel storage area networks when failures in links occur, quantifying how much the global SAN performance is reduced during the time the system remains in the degraded state. We perform this analysis by using both synthetic and real I/O traffic. Simulation results show that performance degradation mainly depends on the routing algorithm and the switch architecture used.

Performance analysis of storage area networks using high-speed LAN interconnects

Storage area networks (SANs) are an emerging data communications platform which interconnects servers an storage devices (such as disks, disk arrays, and tape drives) to create a pool of storage that users can access directly. SANs eliminate the bandwidth bottlenecks and scalability limitations imposed by previous SCSI bus-based architectures and LAN connections between servers and the stored data. This networking approach reports benefits such as computer clustering, topological flexibility, fault tolerance, high availability, and remote management. The prominent technology for implementing SANs is the fibre channel, due to the suitability of this technology for storage networking. Other technologies for high performance interconnects have also been developed. These interconnects provide switch-based networks with links transferring data at more than 1 Gigabit per second, being mainly used in the LAN environments. We analyze whether these high-speed LAN technologies could also be an interesting alternative to storage networking. We perform this analysis using real-world I/O traces. The main conclusion from our study is that most of the messages present the base network latency, meaning that the network is not heavily loaded. Moreover the response time is, in general, acceptable, being dominated by the time disks need to process the requests.

A tool for the design and evaluation of fibre channel storage area networks

The fast growth of data intensive applications has caused a change in the traditional storage model. The server-to-disk approach, usually implemented with SCSI buses, is being replaced by storage area networks (SAN), which enable storage to be externalized from servers, thus allowing storage devices to be shared among multiple servers. A SAN is a separate network for storage, isolated from the messaging network and optimized for the movement of data between servers and storage devices. Nowadays, most current SAN use Fibre Channel as the technology to move data between servers and storage devices. In order to design and evaluate the performance of these systems it is necessary to have adequate tools. Usually, performance evaluation may be based on analytical modeling or simulation. Each of them differs in their scope and applicability or simulation modeling technique offers more freedom, flexibility, and accuracy than analytical methods. Thus, when evaluating the performance of SAN, simulation modeling should be used. We present the main capabilities of a simulator for Fibre Channel SAN, focusing on its input parameters and output variables. We also show several simple examples of performance measurements that can be obtained using this tool.