Now showing items 20-23 of 23
| Abstract: | Existing networked filesystems are usually either client/server - allowing storage only on one node - hard to use, or both. The advanced ones also like to use their own on-disk format, complicating migration both ways. Skynet attempts to remedy this. It is a distributed filesystem with master/slave redundancy that is easy to use, relatively safe for your data and can be easily converted to/from a non-distributed filesystem. It uses an existing filesystem for file storage, takes care of its own maintenance as far as possible, and supports approximate POSIX semantics, with POSIX, eventual and session coherency modes. It favors speed over correctness, where this would rarely be noticed. Part of the project includes a cryptographically secure message-passing middleware for Haskell called Hermes, with distribution transparency and a gossip system. High-level SHA2 and AES bindings are also included. The Hermes portion of the project is complete and usable, the Skynet portion is not. Skynet and Hermes are designed for small-to-medium networks, achieving optimal performance at this size and suffering significant degradation in large networks. They are meant to be used for non-administrated applications in home networks. Skynet offers single-bit security. The network is encrypted and authenticated, but there is no further security inside the network. |
| URI: | http://hdl.handle.net/10037/2543 |
| Abstract: | This dissertation studies composing a super sensor network from the combination of three functional sensor networks; A Sensor data producing network, a sensor data computing network and a sensor controlling network. The target devices are today labeled as large sensor nodes. The communication are based on an IP network using HTTP as the main protocol. Bonjour is used for service discovery, with some adjustments for technical reasons. This allows for naming and location of available services without centralized servers, and it is implementable in small devices. A super sensor network for meteorological observations is emulated using a computer cluster. The emulated measurements are accessed from stations available from observation collection systems accessible on the Internet. Images from web cameras are one kind of observation type used. The implemented system uses Python for rapid prototyping and for support for multiple operating systems. This dissertation demonstrates that the selected technology and architecture may handle some of the demands in a sensor network, and that the architecture gives new opportunities on how to handle updates and sensor network control. The implemented system also demonstrates that using standard Internet protocols can make access to services in the sensor network easy. A web browser may become the preferred user interface for controlling and accessing all parts of the sensor network, as it has for controlling printers and simple network devices. |
| URI: | http://hdl.handle.net/10037/1445 |
| Abstract: | I dag er elektronisk post en naturlig måte å kommunisere på. Dessverre er det enkelte brukergrupper som ikke kan utnytte denne teknologien. Vårt fokus er barn i alderen 4-8 år. Symbo prosjektet tar sikte på å gi disse brukerne muligheten til nettopp dette. Målet med dette prosjektet er å lage et symbolbasert språk kalt SymboL. Dette er en samling symboler som gjennom bruk gir et begrenset ordforråd, setningsoppbygning og grammatikk. SymboL skal så benyttes i en applikasjon kalt Symbo, som sender og mottar symbolbasert e-post. I utviklingen av SymboL har vi samarbeidet med barn i alderen 4-8 år. Vi har drevet med lavteknologisk prototypearbeid, for å få de data som ga oss våre kriterier til design av SymboL. Barna har bidratt med informasjon om hvordan symbolene skulle se ut, og hvordan de skulle grupperes. Etter design og implementasjon, ble SymboL testet sammen med applikasjonen Symbo. Barna var også her med og evaluerte både SymboL og Symbo. Det var helt nødvendig for oss å bruke publiserte metoder for utvikling av programvare til barn. Blant annet var bruk av videokamera for å dokumentere vårt arbeid uunnværlig. Vi gjennomførte vårt mål med å gi barna et symbolbasert E-post system der de kunne kommunisere med andre. Barna både likte og mestret denne formen for kommunikasjon. Det var liten tid til å teste SymboL språket skikkelig, men vi fikk indikasjoner om at SymboL kan forbedres. |
| Description: | Dette er en hovedoppgave |
| URI: | http://hdl.handle.net/10037/1043 |
| Abstract: | To achieve low overhead, traditional cluster monitoring systems sample data at low frequencies and with coarse granularity. However, interactive monitoring requires frequent (up to 60 Hz) sampling of fine-grained data and visualization tools that can explore and display data in near real-time. This makes traditional cluster monitoring systems unsuited for interactive monitoring of distributed cluster applications, as they fail to capture short-duration events, making understanding the performance relationship between processes on the same or different nodes difficult. To address this issue, WallMon was developed, a tool for interactive visual exploration of performance behaviors in distributed systems. For gathering of data, WallMon is centered around an abstraction of collectors and handlers; collectors gathers data of interest, such as CPU and memory usage, and forwards it to handlers in a push-based fashion, while handlers take action upon the data. WallMon captures and visualizes data for every process on every node, as well as overall node statistics. Data is visualized using a technique inspired by the concept of information flocking. WallMon's design is based on the client-server model, and it is extensible through a module system that encapsulates functionality specific to monitoring (collectors) and visualization (handlers). A set of experiments have been carried out on a cluster of 29 nodes with 180 processes per node. Performance results show 7% (of 100) CPU usage at 64 Hz sampling rate when performing process-level monitoring with WallMon. Using WallMon's interactive visualization, we have observed interesting patterns in different parallel and distributed systems, such as unexpected ratio of user- and kernel-level execution among processes in a particular distributed system. |
| URI: | http://hdl.handle.net/10037/3991 |
Now showing items 20-23 of 23
Munin is powered by DSpace 1.8.2
The University Library of Tromsø, N-9037 Tromsø
Tel: +47 77 64 40 00, E-mail: munin@ub.uit.no