Research focus of the XtremWeb project

The main application domain of the Large Scale Distributed System developed in Grand-Large is high performance computing. The two main programming models associated with our platform (RPC and MPI) allow to program a large variety of distributed/parallel algorithms following computational paradigms like bag of tasks, parameter sweep, workflow, dataflow, master worker, recursive exploration with RPC, and SPMD with MPI. The RPC programming model can be used to execute concurrently different applications codes, the same application code with different parameters and library function codes. In all these cases, there is no need to change the code. The code must only be compiled for the target execution environment. LSDS are particularly useful for users having large computational needs. They could typically be used in Research and Development departments of Pharmacology, Aerospace, Automotive, Electronics, Petroleum, Energy, Meteorology industries. LSDS can also be used for other purposes than CPU intensive applications. Other resources of the connected PCs can be used like their memory, disc space and networking capacities. A Large Scale Distributed System like XtremWeb can typically be used to harness and coordinated the usage of these resources. In that case XtremWeb deploys on Workers services dedicated to provide and manage a disc space and the network connection. The storage service can be used for large scale distributed fault tolerant storage and distributed storage of very large files. The networking service can be used for server tests in real life conditions (workers deployed on Internet are coordinated to stress a web server) and for networking infrastructure tests in real like conditions (workers of known characteristics are coordinated to stress the network infrastructure between them).

Application data movements and storage are major issues of LSDS since a large class of computing applications requires the access of large data sets as input parameters, intermediary results or output results.

Several architectures exist for application parameters and results communication between the client node and the computing ones. XtremWeb uses an indirect transfer through the task scheduler which is implemented by a middle tier between client and computing nodes. When a client submits a task, it encompasses the application parameters in the task request message. When a computing node terminates a task, it transfers it to the middle tier. The client can then collect the task results from the middle tier. BOINC follows a different architecture using a data server as intermediary node between the client and the computing nodes. All data transfers still pass through a middle tier (the data server). DataSynapse (http://www.datasynapse.com) allows direct communications between the client and computing nodes. This architecture is close to the one of file sharing P2P systems. The client uploads the parameters to the selected computing nodes which return the task results using the same channel. Ultimately, the system should be able to select the appropriate transfer approach according to the performance and fault tolerance issues. We will use real deployments of XtremWeb to compare the merits of these approaches.

Currently there is no LSDS system dedicated to computing that allows the persistent storage of data in the participating nodes. Several LSDS systems dedicated to data storage are emerging such as OCEAN Store [77] and Ocean [62]. Storing large data sets on volatile nodes requires replication techniques. In CAN and Freenet, the documents are stored in a single piece. In OceanStore, Fastrack and eDonkey, the participants store segments of documents. This allows segment replications and the simultaneous transfer of several documents segments. In the CGP2P project, a storage system called US has been proposed. It relies on the notion of blocs (well known in hard disc drivers). Redundancy techniques complement the mechanisms and provide raid like properties for fault tolerance. We will evaluate the different proposed approaches and the how replication, affinity, cache and persistence influence the performances of computational demanding applications.

Scheduling is one of the system fundamental mechanisms. Several studies have been conducted in the context of Grid mostly considering bag of tasks, parameter sweep or workflow applications [60] , [58]. Recently some researches consider scheduling and migrating MPI applications on Grid [91]. Other related researches concern scheduling for cycle stealing environments [86]. Some of these studies consider not only the dynamic CPU workload but also the network occupation and performance as basis for scheduling decisions. They often refer to NWS which is a fundamental component for discovering the dynamic parameters of a Grid. There are very few researches in the context of LSDS and no existing practical ways to measure the workload dynamics of each component of the system (NWS is not scalable). There are several strategies to deal with large scale system: introducing hierarchy or/and giving more autonomy to the nodes of the distributed system. The purpose of this research is to evaluate the benefit of these two strategies in the context of LSDS where nodes are volatile. In particular we are studying algorithms for fully distributed and asynchronous scheduling, where nodes take scheduling decisions only based on local parameters and information coming from their direct neighbors in the system topology. In order to understand the phenomena related to full distribution, asynchrony and volatility, we are building a simulation framework called V-Grid. This framework, based on the Swarm [83] multi-agent simulator, allows describing an algorithm, simulating its execution by thousands of nodes and visualizing dynamically the evolution of parameters, the distribution of tasks among the nodes in a 2D representation and the dynamics of the system with a 3D representation. We believe that visualization and experimentation are a first necessary step before any formalization since we first need to understand the fundamental characteristics of the systems before being able to model them.