RTDB: A memory resident real-time object database Page: 3 of 3
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to UNT Digital Library by the UNT Libraries Government Documents Department.
Extracted Text
The following text was automatically extracted from the image on this page using optical character recognition software:
V. APPLICATION OF RTDB
The RTDB system has been used in the implementation of
the Distributed Monitoring and Control System (DMCS) [6],
[7]. The DMCS serves a dual purpose; it is required to
monitor and control the environment while performing
special tests and measurements of accelerator magnets.
Therefore, the system has been constructed as a general-
purpose monitoring and control system and a measurement
tool at the same time.
The DMCS is a multi-platform system consisting of a set
of distributed computers connected via a local area network.
Functionally, the system follows the traditional, hierarchical
approach: graphical user interfaces and data analysis
applications run on workstations under Unix while data
acquisition and direct control run on process control
computers under the VxWorks real-time operating system.
Most of the system components are available on both
platforms. System processes, distributed between multiple
nodes, communicate via a software bus.
The system configuration is described using the Data Base
Definition Language (DBDL). Currently, the system
description includes four classes of objects: scans, process
variables, devices, and calculations. DBDL enables the user
to define all of the objects present in the system. The
resulting configuration is loaded to the process control
computers for interpretation. All configuration objects are
kept in RTDB.
RTDB plays a central role in the system. It resides in each
of the process control computers and stores the system
configuration, the state of the process under control, test
conditions, and measurement results.
The RTDB query mechanism is used primarily to browse
through stored objects and select objects of interest to the
user via the provided graphical user interfaces.
The trigger mechanism has been used to implement alarms.
Upon the modification of an object, a trigger handler
procedure examines the object's state and determines if any of
the defined alarm values has been exceeded. If it is the case,
an alarm event is broadcast.
The trigger mechanism has been also used to implement
the software quench detection for the high temperature
superconducting leads [9]. Voltages and temperatures along
the leads have been defined as quench thresholds. Exceeding
one of the defined thresholds is equivalent to detecting a
quench, and this results in a request sent to the power control
system [8] to ramp down the current. The voltage and
temperature values are checked in a trigger handler executed
whenever any of the temperatures or voltages is updated in
RTDB by a monitoring scan.
The DMCS databases contain a practically constant set of
objects and have a well-defined phase of creation, followed
only by accesses and modifications, but no insertions.
Therefore, instead of chaining, overflow bucket creation or
rehashing, the hash table is reorganized, which offers a very
good performance for this application.The dynamic contents of the databases are continuously
written to the non-volatile memory by specialized data
archival scans.
VI. CONCLUSION
Memory-resident databases offer an interesting
architectural alternative when building multi-process systems.
Thanks to short and predictable access times, they seem to be
especially suited for real-time applications and data
acquisition systems. Moreover, in-memory databases can,
similarly to traditional databases, be constructed as
distributed systems. As repositories to be primarily accessed
programmatically, they can follow an object model rather
than a relational model of data.
RTDB, being a toolkit, offers many opportunities for
extending and improving. One of the possible development
directions is to work on increased robustness of the system. It
could be achieved, for example, by monitoring processes to
detect process exceptions resulting in blocking the database,
so the whole system is not brought to a halt by an errant
process. This problem is similar to a classical deadlock
caused by a process suspended while inside a critical region.
Memory-resident databases can also be used in multilevel
database systems and form, together with object persistent
storage systems and relational databases, hierarchical
database storage systems [10].
VII. ACKNOWLEDGEMENT
Authors would like to thank John Tompkins and Mike
Lamm for their support and for making this work possible.
Jerzy Nogiec thanks Sergey Sharonov for his friendly
remarks during the development of RTDB and Kelley
Trombly-Freytag for valuable comments on the text of this
article.
VIII. REFERENCES
[1] http://www.mcobject.com/extremedb.htm
[2] http://www.solidteeh.com
[3] http://www.polyhedra.com
[4] http://www.powerdata.com
[5] http://www.vista-control.com
[6] J. M. Nogiec, E. Desavouret, D. Orris, J. Pachnik, S. Sharonov, J. Sim,
J. C. Tompkins, and K. Trombly-Freytag, "A Distributed Monitoring
and Control System," International Particle Accelerator Conference
PAC'97, Vancouver, 1997
[7] J. M. Nogiec, E. Desavouret, J. Pachnik, S. Sharonov, and J. Sim, "An
Open Distributed Monitoring and Control System," Proceedings of the
International Conference on Computing in High Energy Physics
CHEP'97, Berlin, 1997
[8] J. M. Nogiec, E. Desavouret, "Distributed Power Supply Control and
Monitoring System," ICALEPCS 2001, San Jose, 2001
[9] J. M. Nogiec, S. Feher, D. F. Orris, J. Sim, M. Tartaglia, "Architecture
of HTS Software Protection System," International Particle Accelerator
Conference PAC'99, New York, 1999
[10] J. M. Nogiec, K. Trombly-Freytag, D. Walbridge, "Hierarchical Data
Archival System for EMS," PCaPAC 2002, Frascati, Italy, 2002
Search Inside
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Nogiec, Jerzy M. & Desavouret, Eugene. RTDB: A memory resident real-time object database, article, June 4, 2003; Batavia, Illinois. (https://digital.library.unt.edu/ark:/67531/metadc739851/m1/3/: accessed April 17, 2024), University of North Texas Libraries, UNT Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.