|
|
 |
The impetus behind the formation of ESnet developed in the mid-1980s,
when both the Fusion Energy (FE) and High Energy Physics (HEP) programs
recognized the need for substantially improved computer network
facilities. Until then, the Fusion Energy Community had been served
by MFEnet, which was launched in 1976 as a result of the opening
of a dedicated Fusion Energy supercomputer center at Lawrence Livermore
National Laboratory (LLNL) in 1974. In order to make use of the
new National Magnetic Fusion Energy Computer Center (NMFECC), FE
researchers needed high-speed data links between their home sites
and LLNL. This need was met by the initial MFEnet configuration,
in which satellite links connected LLNL to a handful of key national
laboratories and numerous tail circuits linked those labs to the
other FE sites. By the mid-1980s, MFEnet had evolved from a medium
for access to the NMFECC into a general-purpose network for Magnetic
Fusion researchers.
HEP researchers had begun to use computer networking as soon
as it became practical to do so, in the late 1970s. These first
efforts involved microwave links between the Stanford Linear Accelerator
Center (SLAC) and Lawrence Berkeley Laboratory (LBL). In the early
1980s a satellite link was established between SLAC and Argonne
National Laboratory (ANL) to support a HEP experiment at SLAC.
By the mid-1980s, the HEP program had developed an extensive network
of leased lines (mostly operating at 9600 bps) that interconnected
the main particle-accelerator laboratories with numerous other
sites. Until that time, ad hoc network management by volunteers
from the HEP community had served the de facto HEPnet well. However,
this system was expected to encounter serious difficulties managing
the substantial upgrades that had become imperative as HEPnet
utilization began to extend beyond the HEP community.
The early 1980s also saw other ER research programs joining
established computer networks. Many university research groups
began to use the electronic mail and file transfer facilities
of BITnet or ARPAnet to communicate with their collaborators at
the national laboratories. Other university groups found it necessary
to lease direct connections to mainframe computers located at
remote laboratories where those groups' research activities were
concentrated.
The need for a more comprehensive approach to ER networking
began to be appreciated in 1985. In that year, the HEP program
established the HEPnet Technical Coordinating Committee to address
HEPnet's intensifying management needs. In the same year, a subpanel
of the DOE's High Energy Physics Advisory Panel (HEPAP) recommended
the establishment of a formal HEPnet backbone to provide more
effective and efficient networking for the HEP community.[1] This
backbone was to consist of high-speed (56 kbps) trunk lines connecting
the major HEP laboratories. This recommendation coincided in time
with a proposal to upgrade the MFEnet.
Later in FY 1985, Dr. Alvin Trivelpiece, then Director of Energy
Research, charged OER's Scientific Computing Staff (now the Office
of Scientific Computing) with surveying computer networking requirements
across all the ER programs and evaluating the status of existing
network facilities. The results of this survey demonstrated that
enhanced networking facilities were needed to improve access to
unique ER scientific facilities, to facilitate the dissemination
of information among scientific collaborators throughout all the
ER programs, and to expand access to existing supercomputer facilities.
As a result of these results, Dr. Trivelpiece recommended that
the MFEnet and HEPnet initiatives be combined into what would
become ESnet, in order to optimize the efficiency and functionality
of ER-wide networking. In a special presentation made in response
to the survey's findings, the SCS staff set forth a number of
more specific recommendations that became the foundations of ESnet.
The staff recommended the formation of the Energy Sciences Network
Steering Committee to represent the ER scientific community. The
SCS staff also proposed an evolutionary model for the development
of the new network and endorsed a phased approach to achieving
long-term networking goals.
In October 1986, Dr. Trivelpiece approved a formal plan for
the establishment of ESnet. The overall goal of the initiative
was to create a single general-purpose scientific network for
the ER community. The basic approach taken in organizing the new
network was to combine the various ER programs' network activities
by coordinating the applications-level requirements through the
SCS staff. The ESnet Steering Committee was to be a source of
guidance concerning these requirements and a source of general
strategic oversight. The installation, coordination, and day-to-day
operation of ESnet was to be the responsibility of the staff of
the NMFECC, which had been renamed the National Energy Research
Supercomputer Center (NERSC) to reflect its expanded role of providing
supercomputer access and network services to a wider community.
The ESnet Steering Committee held its first meetings in late
1986, and ESnet began providing ER-wide networking services in
January 1988. Initially, ESnet used time-division multiplexing
across X.25 backbone lines, which operated at speeds of 56 kbps
and 256 kbps. By 1989, ESnet had begun deployment of commercially
supplied multiprotocol routers and T1 backbone lines, which provided
speeds of 1.5 Mbps. This configuration became fully operational
in early 1990, with 19 major OER-supported sites connected to
the backbone. Although the performance level supported by the
T1 backbone seemed quite respectable at first, by early 1991 it
was becoming clear that a bandwidth upgrade would be required
within a very few years, and planning for T3 (45 Mbps) capability
began almost immediately.
During the late 1980s and the early 1990s, ESnet also began
providing international connectivity in support of ER program
activities. International connectivity was shared and coordinated
with NASA, NSF (the National Science Foundation), and DARPA (the
Defense Advanced Research Projects Agency), an approach that established
a framework for future interagency cooperation. In the same time
period, ESnet began connecting to regional NSF networks, thereby
providing more ubiquitous network communications through which
university researchers could utilize DOE/ER laboratories and facilities.
Today, as an integral part of the Internet, the Energy Sciences
Network provides seamless, multiprotocol connectivity among a
variety of scientific facilities and computing resources in support
of collaborative research, both nationwide and internationally.
ESnet also supports DOE-sponsored educational activities.
High-performance computing has now become a critical tool for
scientific and engineering research. In many fields of research,
computational science and engineering have become as important
as the more traditional methods of theory and experiment. Progress
and productivity in such fields depend on interactions between
people and machines located at widely dispersed sites, interactions
that can only occur rapidly enough via the medium of high-performance
computer networks. The ubiquity of networks has provided researchers
with unexpected capabilities and unique opportunities for collaborations.
These benefits have only whetted the scientific community's
appetite for still higher levels of networks performance to support
wider network usage, the transmission of ever-greater volumes
of information at faster rates, and the use of more sophisticated
applications. The scientific community is also increasingly sensitive
to the importance of protecting privacy and intellectual property.
The mission of the Energy Sciences Network is to satisfy these
needs as fully as possible for Department of Energy researchers.
Throughout the evolution of ESnet, its managers and the members
of its committees have made significant contributions to the development
of the worldwide Internet. Today the ESnet Steering Committee
is keenly aware of the importance and scope of the proposed National
Information Infrastructure and is formulating appropriate plans
for OER's participation in NII development.
- J. Ballam et al., "Computing for Particle Physics,"
DOE/ER-0234, August 1985.
Go to the next section, ESnet Processes
Go to the preceding section, Introduction
Go to the Table of Contents
|
|
