CMS monitoring with Web services

The main idea

The main idea is to build a Virtual Counting Room based on Web services.

Glossary

At this point a glossary is absolutely necessary otherwise we get lost.

Ant - Program which has the same functionality of make but with Makefile replaced by XML.
APROM - CMS Analysis project.
ARDA - Project for Distributed Analysis for LHC
Clarens - Open source project providing software (in python for the Apache python module) to create portals of Grid services to be used for example for distributed analysis.
Dataset(DS) - Any sort of collection of information.If registered in the Grid they are write-once read-many objects.A dataset is registered with a unique LDN(Logical Dataset Name).The DS is uploaded to the Grid the first time and then can be replicated to a SE any number of times.
Data grids - Grids that provide shared data storage.Based on a Catalog where Logical File Names are associated to Phisical File Names.
DCS - Part of Cms online system
EDG - EU DataGrid project . Its middleware implements a Grid where each node may have one or more of these roles: RB(Resource Broker), IS(Information Server), RM(Replica Manager), RC(Replica Catalog), CE(Computing Element) which controls one or more WN(Worker Node), SE(Storage Element), UE(User Interface) user access node, etc ...
EGEE - Enabling Grids for E-science in Europe. European project that continues the work done by LCG of producing Grid middleware. LCG-2 is EGEE-0 EGEE-1 will be based on Web Services.
Filter Farm - Part of Cms online system Farm . Is composed by FU and runs COSINE.
GENIUS - A Grid Portal set up by INFN to facilitate Web access to Grid applications.
GGF - Global Grid Forum : produces specifications for the grid.OGSI and OGSA are produced by this forum.GGF is for the Grid what W3C is for the Web.
GILDA - Suite of grid elements and applications dedicated to dissemination purposes.In practice a Grid demonstration set up by INFN in Italy.
gLite - Middleware based on WS1 provided by EGEE as a first approximation waiting for WSRF to complete its definition of standard protocols. Its focus is on data analysis (it is a first implementation of ARDA "blueprint") . It will exist for some time with LCG-2 which is more focused on production. In the end they must become a single middleware.
Globus Toolkit - Open Source software widely used to develop Grid middleware. Its main components are GRAM(Resources Management), MDS based on LDAP(Information Service), GridFTP(File Transfer), GSI(Security). Globus is also behind the WSRF approach to Web services.
GT3 - Globus Toolkit version 3 : an open source implementation of OGSI.
HEPCAL - High Energy Physics Application Layer : Analysis of user requirements for LHC experiments done with use case technique.In these use cases we have a definition of : Datasets, Catalogs, Jobs, Job Catalogs, Event, VO, etc...
LCG - LHC Computing Grid project. Has produced Grid Middleware based on GT2. Its work is continued by EGEE.
middleware - Software layer that implements the Grid.Its name comes from the fact that it is in the middle between the application layer and the OS and Network layer.
MONARC - Distributed Computing Model for HEP based on a hyerarchical structure of nodes (Tier0-4)
OGSA - Open Grid Services Architecture: GGF specification to provide interoperability of grid services.
OGSI - Open Grid Services Infrastructure: specification by GGF that describes how standard web services can be used to support grid services.
Realtime Grid - Grid with enough resources to control interactively in realtime remote instruments.
RCMS - Run Control and Monitor System for CMS Experiment
RLS - Replica Location Service : part of LDR (Lightweight Data Replicator) keeps track of the locations of the files within an organization. This is a tool provided by Globus.
Resources - The Grid allows distributed resources management. It makes virtual resources available to Grid users.The main two resources are CPU(computational grids) and data storage(data grids).But other resources can be made virtual.For example access to detectors(virtual counting room).
Security Infrastructure - The Grid has a security infrastucture based on PKI and CA certificates. Normally Grid users can access only the resources of their VO.Inside a VO there are users with different privileges.Each user must have a X.509 certificate issued by a Certification Authority(CA).
Services - The Grid provides virtual services.These are made available by service provider through an information service. A resource broker service is used to catalog available services and make easier their discovery.
SOAP - Protocol used for communication in Web services.
SRM - Storage Resource Management: Castor and dCache are examples of SRM implementations.
XDAQ - Framework for CMS data acquisition.
Telepresence - One of the services provided by a interactive grid used to control remote instruments.
VO - Virtual Organization : Grid users are grouped in VO.VO are grid service providers and have also an important role in the security infrastructure .
XDAQ - Part of Cms online system
XML - data format for data exchange.
Web Services(WS) - Web applications based on four protocols that provide data represesentation(XML), communication (SOAP),description(WDSL) and discovery(UDDI). All these are platform-independent technologies defined as W3C standard and allowing programs to access your application.
WSRF - Web Services Resource Framework: set of six web services specifications . The main reason for these is to have now the possibility to develop Grid services based on GT (GGF is too slow). The six specifications are: ResourceLifetime, ResourceProperties, Notification, RenewableReference, ServiceGroup,BaseFaults.

Web services, Grids and XML

Web services in our context means simply that we want our monitoring services to be accessible from Internet using its protocols. Although this ,in some cases, may require only writing a simple CGI-script or a PHP page backed by a Mysql database, in general we need the possibility to grant users the full access to powerful computing facilities , to online instruments control, etc This kind of access requires in fact the implementation of Grids.The best choice now for implementing computing grids is the Globus Toolkit at Version 3 that implements OGSI 1.1 i.e. Grid enabled Web Services defined by GGF.

Testing the Virtual Control Room with a Tracker display program

The visualization software group in Bari has developed a Java program that is able to display all the modules of the tracker with their features (including the signal) coded in some graphical way. This program is now used to display data stored in the construction database through an interface based on XML.
We could use the same program to put to test the idea of interactive grid.
GRIDCC(Grid-enabled Remote Instruments with Distributed Control and Computation) application proposal
Monitoring CMS Tracker construction and data quality with a grid/web service based on a visualisation tool (draft for CHEP04)

Java Grid Services with GT3

GT3 implements OGSI 1.1 and has a Java interface.Creating Grid Services on a Linux machine requires:

Download and install GT3
Create the service interface using normal Java code.From this code a WDSL description is authomatically generated.
Generate stub and other code by running a special program
Implement the service in Java
Deploy the service
To access the service a client can use Soap or JAX-RPC.

Bibliography

Maintained by Giuseppe Zito:Giuseppe.Zito@cern.ch.Last revised