Andrzej Kochut

Application Performance Management in Virtualized Server Environments

As businesses have grown, so has the need to deploy I/T applications rapidly to support the expanding business processes. Often, this growth was achieved in an unplanned way: each time a new application was needed a new server along with the application software was deployed and new storage elements were purchased. In many cases this has led to what is often referred to as server sprawl, resulting in low server utilization and high system management costs. An architectural approach that is becoming increasingly popular to address this problem is known as server virtualization. In this paper we introduce the concept of server consolidation using virtualization and point out associated issues that arise in the area of application performance. We show how some of these problems can be solved by monitoring key performance metrics and using the data to trigger migration of virtual machines within physical servers. The algorithms we present attempt to minimize the cost of migration and maintain acceptable application performance levels

Sniffing out the correct physical layer capture model in 802.11b

Physical layer capture (PLC) in 802.11b refers to the successful reception of the stronger (higher signal strength at receiver) frame in a collision. PLC causes significant imbalance in the throughputs of sources. Existing 802.11b simulators, including ns2 and Qualnet, assume that PLC occurs only if the stronger frame arrives first at the receiver. We show empirically that in reality PLC occurs even if the stronger frame arrives later (but within the physical layer preamble of the first frame). Consequently, throughput unfairness in reality can be significantly (up to 15%) higher than with the former PLC model. We have modified the ns2 simulator to account for this and Qualnet is incorporating a fix in their next release. To identify which frames were involved in collisions, when their transmissions started, and which of them were retrieved, we have devised a novel technique using multiple sniffers and instrumented device drivers to reconstruct from the air interface all tx/rx events in a WLAN to within 4 /spl mu/s accuracy. This allows us to quantify the causal links from the PHY layer through the MAC layer to the observed application layer imbalance. It also shows that the arrival times of colliding frames routinely differ by as much as 20 /spl mu/s due to inherent uncertainties of 802.11b firmware clock synchronization and rx/tx turnaround delays, and that the frame to arrive first can be either the stronger or the weaker with equal likelihood.

Rover: scalable location-aware computing

All the components necessary for realizing location-aware computing are available in the marketplace today. What has hindered the widespread deployment of location-based systems is the lack of an integration architecture that scales with user populations. The authors have completed the initial implementation of Rover, a system designed to achieve this sort of integration and to automatically tailor information and services to a mobile users location. Their studies have validated Rovers underlying software architecture, which achieves system scalability through high-resolution, application-specific resource scheduling at the servers and network. The authors believe that this technology will greatly enhance the user experience in many places, including museums, amusement and theme parks, shopping malls, game fields, offices, and business centers. They designed the system specifically to scale to large user populations and expect its benefits to increase with them.

Octree approximation an compression methods

Techniques are presented to progressively approximate and compress in a lossless manner two-colored (i.e. binary) 3D objects (as well as objects of arbitrary dimensionality). The objects are represented by a region octree implemented using a pointerless representation based on locational codes. Approximation is achieved through the use of a forest. This method labels the internal nodes of the octree as GB or GW, depending on the number of children being of type GB or GW. In addition, all BLACK nodes are labeled GB, while all WHITE nodes are labeled GW. A number of different image approximation methods are discussed that make use of a forest. The advantage of these methods is that they are progressive which means that as more of the object is transmitted, the better is the approximation. This makes these methods attractive for use on the worldwide web. Progressive transmission has the drawback that there is an overhead in requiring extra storage. Aprogressive forest-based approximation and transmission method is presented where the total amount of data that is transmitted is not larger than MIN(B,W), where B and W are the number of BLACK and WHITE blocks, respectively, in the region octree of the set of objects.

Managing virtual storage systems: an approach using dependency analysis

We present an approach for managing the performance of virtual storage systems by experimentally identifying the dependencies that exist between various components that comprise the system. Specifically, we show how one may profile dependencies between each logical volume exported by a storage system and components that this volume uses. To do so the technique estimates the arrival rate and size of requests issued to the internal system component as a functions of arrival rate and size of requests issued to the logical volume. The complete dependency profile of the system consists of a set of such functions for READ and WRITE operations separately and for each pair: logical volume - internal system component. The empirical technique of obtaining such profiles for typical existing storage systems is presented. We propose the use of Common Information Model (CIM) as a way to express dependency and performance information in an architecture-independent manner. The dependencies between components are computed as a fraction of bandwidth that is passed on to the sub-components. We discuss how the dependency profile of the system may be used to perform root-cause analysis and early Service Level Agreement violation notification. We also demonstrate the use of the method by applying it to a Linux system using software RAID.

Programming computers embedded in the physical world

During the next decade, emerging technologies will help populating the physical space with ubiquitous networks of embedded systems (NES). Programming NES requires new abstractions and computing models since the current programming, models are not designed for the scale and volatility encountered in these networks. This paper presents Spatial Programming (SP), a novel programming model for NES. The key idea in SP is to offer network-transparent access to data and services distributed on nodes spread across the physical space. SP programs access network resources using a high level abstraction, termed spatial reference, which addresses the nodes using their spatial and content properties. An underlying system takes care of mapping spatial references onto target nodes in the network. Our preliminary experience in developing SP applications suggests that SP can be a viable solution for distributed computing over NES.

Management issues in storage area networks: detection and isolation of performance problems

This paper explores merging SAN performance management tools into the host system thus dissolving the traditional boundary between hosts and storage. The host service time distribution is exploited to locate caches in the SAN, detect and isolate contention in the communication channels and backend disk drives. We define essential metrics and measure I/O workloads over extended periods to define baseline behavior. We have implemented architecture to acquire configuration and performance data from agents on the host and SAN elements. In addition to performance management this data is the basis of management functions such as capacity planning, and sizing host and SAN resources required for application rollout and expansion. The static dependency map (SAN topology or wiring diagram), provided by current SAN management tools, is extended into the host server. We then attempt to develop a corresponding dynamic dependency view showing how performance metrics at the application and host level, such as I/O queuing levels and service times, relate to response times in the SAN. The dynamic view is also an important complement to the static dependency graph because the actual traffic flow may be difficult to deduce from the topology.

Implementation of a Scalable Context-Aware Computing System

Context-aware computing involves the automatic tailoring of information and services based on the current location of the user. In this paper, we describe our experience in implementing Rover, a system that enables location-based services, as well as the traditional time-aware, user-aware and device-aware services. To achieve system scalability to very large client sets, Rover servers are implemented in an “action-based” concurrent software architecture that enables fine-grained application-specific scheduling of tasks. We have demonstrated its feasibility through implementations for both outdoor and indoor environments on multiple platforms.