To read the full-text of this research, you can request a copy directly from the author.
... As shown in Fig. 1, MAGNeT differs from tcpdump-like measurement tools in that it can monitor traffic at the application level, (i.e., before it traverses the protocol stack), and throughout the entire protocol stack, as well as traffic entering and leaving the network. The only other measurement tool that makes similar measurements is the TCP kernel monitor from Pittsburgh Supercomputing Center . MAGNeT differs from the TCP kernel monitor in at least three ways. ...
... 1, MAGNeT differs from tcpdump-like measurement tools in that it can monitor traffic at the application level, (i.e., before it traverses the protocol stack), and throughout the entire protocol stack, as well as traffic entering and leaving the network. The only other measurement tool that makes similar measurements is the TCP kernel monitor from Pittsburgh Supercomputing Center . MAGNeT differs from the TCP kernel monitor in at least three ways. ...
Over the last decade, network practitioners have focused on
monitoring, measuring, and characterizing traffic in the network to gain
insight into building critical network components (from the protocol
stack to routers and switches to network interface cards). Previous
research shows that additional insight can be obtained by monitoring
traffic at the application level (i.e., before application-sent traffic
is modulated by the protocol stack) rather than in the network (i.e.,
after it is modulated by the protocol stack). Consequently, this paper
describes a monitor for application-generated network traffic (MAGNeT)
that captures traffic generated by the application rather than traffic
in the network. MAGNeT consists of application programs as well as
modifications to the standard Linux kernel. Together, these tools
provide the capability of monitoring an application's network behavior
and protocol state information in production systems. The use of MAGNeT
will enable the research community to construct a library of real traces
of application-generated traffic from which researchers can more
realistically test network protocol designs and theory. MAGNeT can also
be used to verify the correct operation of protocol enhancements and to
troubleshoot and tune protocol implementations
... It is also possible to modify an application to produce time-stamped logs, or to modify the operating system to record time-stamped significant events. This approach is invasive in that re-writing of the application may be required (Semke, 2000). ...
As distributed computing becomes ubiquitous, the ability to troubleshoot, improve performance, and provide meaningful service level agreements requires some overall understanding of an application's performance from end to end. The ability to drill down to increasing levels of detail as needed would be helpful, preferably with some intelligent guidance as to where to look. Today's efforts to manage end-to-end performance for applications can resemble blind men groping at the unknown. Applications have dif-ferent characteristics, run on platforms with varying power with varying operating systems, and travel over heterogeneous networks. This paper will explain current standards for performance measurement, describe the range of tools available, and identify components that will be useful for analyzing applications running or originating on the campus network at the University of Alabama at Birmingham (UAB). Given the size and complexity of UAB's network, insights gained will be of general interest.
Having the ability of monitoring the operating system kernel is a great feature for a system administrator. This feature makes system management more powerful than ever, i.e by increasing the system security, providing new debugging capabilities or studying the Operating System internals. It is important to take in account that monitoring the kernel involves the manipulation of many internal data structures of the operating system so implementing this feature is only possible for Open Source operating systems, such as Linux. Thus, we have developed a kernel monitor named ESKM (Embedded Space Kernel Monitor) for the Linux kernel. ESKM can freeze the kernel state and show or modify any of the kernel data and internal structures. When ESKM is executed no other process can take the CPU, and interrupts are disabled. ESKM has been developed for a Linux port for an embedded ERC32 system, which is fully SPARC v7 compliant, making it more flexible for debugging and for development. Our main goal is to provide a tool that allows to perform the necessary coverage tests in our system, which is only possible through fail injection. ESKM provides the needed features to do this, so we consider it is also a good tool for embedded system developers.
The current trend in constructing high-performance computing systems is to connect a large number of machines via a fast interconnect or a large-scale network such as the Internet. This approach relies on the performance of the interconnect (or Internet) to enable fast, large-scale distributed computing. A detailed understanding of the communication traffic is required in order to optimize the operation of the entire system.
Network researchers traditionally monitor traffic in the network to gain the insight necessary to optimize network operations. Recent work suggests additional insight can be obtained by also monitoring traffic at the application level.
The Monitor for Application-Generated Network Traffic toolkit (MAGNeT) we describe here monitors application traffic patterns in production systems, thus enabling more highly optimized networks and interconnects for the next generation of high-performance computing systems.
With the growth of high performance networking, a single host mayhave simultaneous connections that vary in bandwidth by as many as six orders of magnitude. We identify requirements for an automatically-tuning TCP to achieve maximum throughput across all connections simultaneously within the resource limits of the sender. Our auto-tuning TCP implementation makes use of several existing technologies and adds dynamically adjusting socket bu#ers to achieve maximum transfer rates on each connection without manual con#guration. Our implementation involved slight modi#cations to a BSD-based socket interface and TCP stack. With these modi#cations, weachieved drastic improvements in performance over large bandwidth*delay paths compared to the default system con#guration, and signi#cant reductions in memory usage compared to hand-tuned connections, allowing servers to support at least twice as many simultaneous connections. 1 Introduction Paths in the Internet span more than 6 orders of magn...
Transmission Control Protocol
[RFC793] "Transmission Control Protocol." IETF RFC 793, September 1981.
Available from: http://www.rfc-editor.org/rfc/rfc793.txt.