SLAC Statement of Work Measurement and Analysis for the Global-grid and PERN’s Internet End-to-end performance (MAGPIE) Year 1: Measurements: Starting immediately, the Measurement Point (MP) services will be developed to provide access to data from active and passive network monitoring tools that are available in PERN2. In addition we will develop endhost MPs for making end-to-end measurements. Following the development of relaibel end-point MPs we will start deployment at major collaborator sites (NIIT, NCP, COMSATS …) proving feedback to improve/streamline the process. Storage: Starting in quarter 2, Measurement Archives (MA) services will be developed and hardened to allow the storage of MP data that will provide historical information regarding the performance of the network. Performance forecasting: starting in quarter 2, we will work on network forecasting using time series data from various MPs. We will implement, test and compare various forecasting techniques to determine their applicability, ease of implementation, ease of understanding and configuration. Year 2: Measurements: in the first quarter we will complete providing the MP s for PERN and complete the first round of collaborator end-site MP deployments. Storage: Work on extending the capabilities, reliability and ease of use of the MAs will continue through quarters 1 through 4. During this time we will start deploying production MAs at PERN and the end-sites Topology: During quarters 2-4 we will develop topology services to provide network information regarding the topology of the network. Towards the end of this period we expect to start deployment of topology services and improve them based on experience and feedback. Discovery: Starting in quarter 2, we will develop the lookup service to provide DNS like mechanisms to discover the virtual location of the various perfSONAR services. Security: Starting in quarter 4, we will choose and integrate suitable AAA services. Network performance forecasting: in quarters 1-3 we will make the forecasting more robust, and accurate. Towards the end of this period the forecasting will be integrated with the production measurements. Anomaly detection: starting in quarter 2, we will start to compare forecasts with actual measured data and develop ways to trigger usable events (detects most important, sustainable events, does not generate false positives) and tune the detection to our needs. Year 3: Topology: for quarter 1 and 2 we will productize the topology services, provide deployment kits and deploy at the appropriate places. Discovery: for quarter 1 and 2, we will productize the topology services, provide deployment kits and deploy at the appropriate places. Security: quarter 1 through 4 we will deploy and integrate the chosen AAA services,. Anomaly Detection and Diagnostic services: in quarter 1 we will explore how to automate the diagnosis of detected anomalies, e.g. by divide and conquer techniques looking at measurements along a path, by looking at data from multiple MPs to look for commonality etc. Integration and Testing: in quarter s 3 and 4 we will provide training, refine the documentation, ensure the software is in a stable state for continued use, and provide needed management tools. Alert Services: in quarters 1 to 4, we will refine the techniques for generating alerts, further diagnosing to reduce false positives, and ways to raise alerts that are acceptable and useful to the network engineers etc. Reporting Service: in quarters 1 through 4 we will develop, integrate and provide management reports on network performance to assist in problem analysis, planning, expectation setting and auditing.