|
Managing Multivendor Networks
  
- 2 -
Digital Equipment Corporation
Company Background
Digital Equipment Corporation (DEC) was founded in 1957 by three graduates of
Massachusetts Institute of Technology (MIT) operating out of an old mill in Maynard,
Massachusetts. One of the original three founders, Ken Olsen, went on to lead the
corporation and has been the most significant person to shape Digital. The fact that
the corporate headquarters remains in Maynard is representative of the company's
paradoxical commitment to both innovation and tradition.
During its early years, Digital produced specialized logic modules. Although it
operated in a data processing world dominated by large, mainframe-style computers,
Digital had the vision to pioneer minicomputers by introducing the PDP-1 in 1960.
The revolutionary aspect of the PDP-1 was its interactive video terminal capability--the
first such implementation of this now commonplace technology.
In 1970, Digital released the PDP-11, a 16-bit machine that became the most popular
design of the PDP series. Used most often in manufacturing and technical environments
for process and instrument control, the PDP-11 family remained in demand until it
was discontinued in 1991. Following the growth of the PDP-11 market, Digital released
a networking architecture in 1976 called the Digital Network Architecture (DNA).
The set of products and services defined by DNA is more commonly known as DECnet.
The PDP-11 design also served as the launching platform for another product: Digital's
general-purpose, 32-bit Virtual Address Extension (VAX) computer. Introduced in 1977,
the VAX gradually became Digital's premiere product--a building block for other products
offering performance and disk capacity similar to that of mainframes. Among these
products are VAX clusters, a computing environment released in 1983 and based on
multiple VAX computers sharing a common repository of disk space; and the mainframe-level
VAX 9000 Series, announced in 1989. In 1992, Digital released the Alpha AXP architecture,
based on 64-bit RISC microprocessors. Today, the Alpha runs on all Digital machines,
from PCs to high-end symmetric multiprocessing and clustered servers. Digital also
offers a line of high-end Pentium Pro products, including the four-processor Prioris
server and Celebris workstation.
Product Line Overview
Digital maintains a comprehensive manufacturing operation that designs and builds
most of its products. Having paved the way for the so-called midrange computer market,
Digital has expanded its product line up into the mainframe domain and down to the
PC playground. In support of these offerings, Digital produces a complete line of
terminals, printers, and assorted networking devices.
Digital's VAX family of computers ranged from the MicroVAX desktop system, all
the way up to high-end symmetric multi-processing (SMP) servers. Every member of
the VAX family supported the same operating systems and applications. The VAX family
was based on CMOS V technology and was highly scalable, up to symmetric multiprocessing
and clustering configurations. A number of different clustering options were possible
under the VAX architecture, including clustering in a single office, or clustering
systems that were geographically separated. However, the 32-bit nature of the VAX
architecture was quickly hitting the wall, primarily because of address limitations.
In 1988, Digital formed a task force to explore ways to preserve its existing
VAX VMS customer base through the coming decade, and in 1992, released the Alpha
AXP architecture. The Alpha AXP architecture retains many of the VAX's attributes
while offering significantly more power. The 32-bit VAX architecture was based on
complex instruction set computing (CISC), whereas the 64-bit Alpha AXP architecture
was reduced instruction set computer (RISC)-based. Digital subsequently ported its
OpenVMS operating system to the Alpha AXP architecture to enable OpenVMS applications
to take advantage of RISC's performance advantages.
The Alpha AXP architecture was designed for high-performance computing. In fact,
OpenVMS AXP applications outperform OpenVMS VAX applications by a factor of 3.59
to 1. Digital built easy migration capabilities to enable customers to move from
the VAX to the Alpha AXP architecture without significant recoding of their applications.
Digital's Alpha AXP processors run multiple operating systems and have the ability
to run native programs translated from VAX and MIPS architectures, thereby preserving
their customers' existing VAX and MIPS investments.
Digital's goals in developing the Alpha AXP architecture were to provide:
- High performance
- Longevity
- The ability to run both VMS and UNIX operating systems
- Easy migration from VAX and MIPS architectures
To operate OpenVMS AXP, DEC OSF/1 AXP, and Microsoft Windows NT operating systems,
Digital adopted some technology from their PRISM design. Under this model, a set
of sub-routines (PALcodes) with controlled entry points were established for each
operating system. For running VAX and MIPS binary images, Digital uses binary translation.
The Alpha AXP architecture is based on a shared-memory model. The first implementation
was the DECchip 21064 microprocessor. At the time of its release in 1992, it was
the world's fastest single-chip microprocessor--even listed in the Guinness Book
of Records as such. The latest implementation, the 21164 Alpha microprocessor,
runs at a blazingly fast 300MHz, and power can be increased further through symmetric
multiprocessing or clustering. The Alpha AXP chip is capable of running multiple
operating systems, and can run native programs translated from VAX and MIPS architectures.
The Alpha AXP architecture is now used throughout Digital's product line. The
Alpha AXP's 64-bit architecture is designed with an eye towards high performance,
and continues Digital's focus on multiple processors. The powerful 64-bit Alpha architecture
is capable of bringing high-end features to smaller systems. The Alpha systems can
run Digital UNIX, OpenVMS, and Windows NT; for sophisticated functions such as data
warehousing, they can address massive files greater than 2G in size.
Digital released a 64-bit version of its Digital UNIX operating system in March,
1996. Digital UNIX 4.0 integrates the now pervasive Common Desktop Environment
(CDE) GUI, which establishes a common look and feel between all major UNIX implementations.
Digital UNIX 4.0 supports POSIX threads, real-time standards, and X11R6. Additionally,
it conforms fully to the Single UNIX Specification (Spec1170) administered by the
X/Open organization. The 64-bit nature of version 4.0 now enables Digital UNIX to
run high-end applications such as data warehousing.
Digital Equipment Terminals
The Digital line of terminals is perhaps the most widely emulated line of character-oriented
displays. The reasons are straightforward:
- Digital does an excellent job of providing ANSI compatibility, which enables
Digital terminals to be used in environments that might not include Digital computer
systems.
- The widespread use of the Digital PDP and VAX computer systems has enabled the
Digital terminal line to penetrate deep into both the technical and end-user communities.
The following generations of products compose the fundamental line of Digital
video terminal (VT) devices:
- VT50 Family. This is the father of the VT line. Although not as capable
as today's products, the VT52 implementation in particular is still widely emulated
by PCs and terminals that require character-oriented access. Its sibling, the VT55,
offered support for combined on-screen text and graphics.
- VT100 Family. Introduced in 1978, the VT100 line offered improved speed
and function over the initial VT50 models. Of significance was the VT100 line's support
for both ANSI and VT52 compatabilities. The original VT100 sired the following models:
- VT101. Became the new low-end product.
- VT102. Offers advanced video options, including support for 132 columns
by 24 lines.
- VT131. Supports both conversational (character-oriented) and block mode
(full-screen) transmission modes.
- VT125. Combines text with bitmapped graphics capabilities. The VT125 includes
implementation of Digital's Remote Graphics Instruction Set (ReGIS).
- VT200 Family. One of the most important features of the VT200 family is
the keyboard layout. While the VT100 had a separate (right-hand) numeric keypad,
the VT200 added another keypad for editing (located between the typewriter keys and
the numeric key-pad). Another significant change from the VT100 family was the introduction
of a user-friendly on-screen menu to define the various set-up options (in contrast
to the cryptic on-screen sequence used by the VT100). The VT200 family is composed
of three major models:
- VT220. A standard monochrome text-only video terminal.
- VT240. Supports text and ReGIS monochrome graphics. It is composed of
a keyboard, monitor, and a system unit box.
- VT241. A color implementation of the VT240.
- VT300 Family. The VT300 line offersincreased performance and ergonomics
(reduced glare and tilt/swivel base) over the VT200 line. The VT300 line includes:
- VT320. The entry-level monochrome text video terminal.
- VT330. The monochrome graphics (ReGIS) replacement for the VT240. The
design of the VT330 eliminated the need for a separate system unit. Also note- worthy
is the VT330's dual-session (dual-port) capability and improved graphics resolution.
An improved, higher performance model of the VT330 was released as the VT330+.
- VT340. A color implementation of the VT330. A high-performance version
of the VT340 was released as the VT340+.
- VT400 Family. The VT400 family was introduced in 1990 as the planned replacement
for the three-year-old VT300 line. The VT400 line includes improved resident fonts
that enable it to display 24, 36, or 48 lines of text per display screen. The VT400
line was introduced with one model, the VT420--a monochrome text terminal. The VT420
supports dual ports (for dual sessions) and provides split-screen and cut-and-paste
functions for managing both sessions.
- VT1000 Family. The VT1000 is a specialized graphics workstation supporting
icons and multiple windows in accordance with the X Window standard. The highly intelligent
VT1000 also provides VT320 terminal emulation and support for both the Local Area
Transport (LAT) protocol, Transmission Control Protocol/Internet Protocol (TCP/IP),
and the standard X Window System protocol.
In most cases, Digital terminals transmit characters as they are pressed on the
keyboard. Issues such as buffering and transmission optimization are handled by the
computer system or by the devices between the terminal and the computer system. These
considerations will be discussed later in this chapter.
PCs
The terms personal computer and Digital Equipment Corporation are
not as synonymous as one might expect. It is not that Digital has been unable to
apply its innovative talents in this arena, or that the resulting products did not
come to market. The real problem rests with timing. Before IBM's release of the first
PC, Digital had also been working on a small, personal microcomputer. At that time
there was no single dominant hardware architecture or dominant operating system on
the market (although the Control Program/Microprocessor [CP/M] was doing well at
that time).
When IBM released the first PC in 1981 (and subsequently set the standards for
today's market), Digital redoubled its efforts to get its own products to market.
Digital produced not one but three possible contenders for the low-end market. Each
had its own distinct advantages and disadvantages, and each was given an opportunity
to prove itself.
The three products that resulted were the DEC Rainbow, the DECmate, and the Professional
300. Of these three, the DEC Rainbow was the only product that came close to being
a clone of the IBM PC; the Rainbow ran Microsoft's MS-DOS operating system but did
not feature the Basic Input/Output System (BIOS) and hardware-level compatibility
with the IBM machines.
In contrast, the DECmate and the Professional 300 lines were corporate-oriented
computers. For example, the DECmate featured sophisticated word-processing capabilities
and could be networked easily with Digital's larger machines. The Professional 300,
on the other hand, was intended to be a desktop version of the Digital's PDP computer.
None of these three products have any inherent, glaring faults, but all three
were released in the shadow of the IBM PC (and the emerging clone market), which
fell on them like a wet wool blanket. Although it is certainly unfair to call any
of the products a market failure, none achieved the celebrity status earned by the
IBM PC. Digital's resigned mindset was further evidenced by the release of the VAXmate,
a follow-up product to the IBM PC/AT. Based on the same Intel 80286 microprocessor
chip as the IBM AT, the VAXmate had limited expansion capabilities and was marketed
only to existing Digital customers.
At the end of the 1980s, Digital was ready to take another chance on the IBM PC-compatible
market. This time, however, it turned to Tandy Corporation (Radio Shack; Fort Worth,
Texas) to manufacture a line of PC-compatible products for Digital. The resulting
products were the DECstation 200, DECstation 300, and DECstation 400 lines--based
on the Intel 80286, 80386, and 80486 processors, respectively. Although Digital downplayed
the fact that it had out-sourced the product to Tandy, this fact was not lost on
the market.
Finally, in 1990, Digital released the applicationDEC 433MP, a multi-user UNIX
system based on the Intel 80486 microprocessor. This product was targeted at the
small-business market and features support for multiple 486 processors and connectivity
for up to 96 concurrent users. Its roots in PC and Personal System/2 (PS/2) technology
are evident by its support for either the Extended Industry Standard Architecture
(EISA) or Micro Channel Architecture (MCA) bus. If nothing else, the 433MP represents
an interesting convergence of PC, engineering workstation, and minicomputer technologies.
Digital's latest innovation, the Digital Personal Workstation, is designed around
a processor-independent architecture, and offers the user a choice between Pentium,
Pentium Pro, or Alpha processors. Users can move from Intel to Alpha with the change
of a single card. Both processors are optimized to run Windows NT, and can be used
to run many of the high-end applications, such as CAD or GIS, that had been limited
to higher-end workstation products. It offers PCI or ISA Ethernet or token ring options
for networking. The Celebris XL is based on Intel technology and features either
single 100, 120, or 133 MHz; or dual 100 or 133 MHz Pentium processors. The Alpha
XL runs 233 MHz and 266 MHz implementations of the Alpha 21064 processor. Several
native Windows NT applications have been ported to the Alpha system as a result of
an alliance between Digital and Microsoft.
Engineering Workstations
Digital's original response to the engineering workstation market was to couple
high-resolution displays with various VAX processors. Capable of running either the
VMS operating system or ULTRIX (Digital's implementation of UNIX), the product line
was termed the VAXstation.
In general, as the capabilities of the mainstream VAX models grew, so did those
of the VAX-stations. The VAXstation 100/500 line introduced in the early 1980s was
replaced by the VAXstation II in 1985, a line that included a model with an independent
graphics coprocessor. On the heels of the VAXstation II came the VAXstation 2000-7000,
released in 1987 and 1988.
The VAX 7000 was designed as a high-performance system, suitable for high-volume
transactions or distributed networks. Many mainframe-based applications could run
on the VAX 7000, making it usable as a backbone platform for supporting business-critical
applications. The VAX 7000, which was expandable to six CPUs, was built on Digital's
CMOS technology and ran the OpenVMS operating system.
In 1989, Digital revamped its product line and stirred some new ingredients into
the mix. The first ingredient was the use of a RISC architecture for some of the
machines. Another was that the basic RISC processor used by Digital was, in fact,
manufactured by MIPS Computer Systems, a third party supplier. Finally, these new
RISC-oriented machines could run only ULTRIX (and thus were shunned from the VAX
community).
Presumably to reduce confusion between the VAX-based and RISC-based products,
Digital named the new line DECstations (a named shared by Digital's MS-DOS
computers). Digital targeted this new lineup as a sweeping, low-end desktop blitz
and paraded it in front of the public.
The product lineups consisted of the VAXstation 3000 Series (based on Digital
complementary metal oxide conductor, or CMOS, VAX processor technology); the
DECstation 5000 Series (based on the MIPS RISC architecture); and the DECstation
200, 300, and 400 Series (based on the Intel 80286, 80386, and 80486 microprocessors).
While the low-end, PC-oriented DECstations were not presented as heirs to the engineering
workstation throne, the association (by name alone) of the DECstation 3000 Series
with these units raised both eyebrows and confusion.
The AlphaStation 255 and 500 families of UNIX and Windows NT workstations are
the latest additions to Digital's line of midrange workstations. Digital's AlphaStations
support Digital's new 64-bit operating systems (Digital UNIX and OpenVMS), as well
as Windows NT.
The 500 supports dual-fast and wide SCSI-2 channels, Ethernet and Fast Ethernet,
and can accommodate up to 512M of RAM. All products in the AlphaStation line include
multimedia capabilities.
The PowerStorm PCI-based workstation graphics option is available for all Digital
workstations. PowerStorm delivers superior graphics performance, using the common
OpenGL API. PowerStorm was designed for 2-D and advanced 3-D applications that require
high performance, such as motion and texture mapping.
Digital's RISC-based AlphaStations offer superior performance for applications
such as modeling, imaging, animation or videoconferencing. The new 64-bit architecture
can directly address up to 1G of real memory, making it easier to handle very large
files without disk-swapping. The newest AlphaStations run the 64-bit RISC Alpha 21164
microprocessor, at speeds up to 300 MHz. Like the other Digital products running
Alpha technology, it can run UNIX, OpenVMS or Windows NT.
The AlphaStation 250 is well-suited to mechanical or electrical CAD applications,
and also offers collaborative computing facilities. The 200 is a somewhat lower-cost,
entry-level product, but also offers the 64-bit computing environment, at speeds
up to 233 MHz. On the high end are the AlphaStation 600 systems, which are the fastest
and most ideal for high-end scientific research projects that involve complex visualization
and calculation.
The AlphaStation 500 (see Figure 2.1) is well-suited for higher-end CAD applications
or memory-intensive and CPU-intensive multimedia projects. The system runs the Alpha
21164 processor at 333, 400, or 500 MHz.
Midrange Offerings
The PDP-11 line, until it was discontinued, filled the lower end of the spectrum,
if for no other reason than by virtue of being a 16-bit machine. The 32-bit VAX,
on the other hand, was offered in a broad range of models, starting from the diminutive
MicroVAX extending to the mainframe-oriented VAX 9000.
FIG. 2.1 Digital Equipment Alpha 500 Workstation
Both lines shared the dubious honor of using multiple bus designs. The three most
common bus architectures are as follows:
- UNIBUS. The architecture used in the original PDP-11 and VAX-11 and at
the high end of other VAX and PDP offerings.
- Q-Bus. Used in the lower end of both the PDP-11 and VAX lines.
- VAXBI. Unlike the Q-bus and Unibus architectures that were open to third-party
vendors for the development of controllers and peripherals, the VAXBI is a closed
architecture. Although this bus structure is specific to the higher end of the VAX
line, it also supports the older Unibus design. Other buses, including the MASSBUS
and the XMI and industry-standard VME bus, have also been used by Digital. Presently,
Digital offers a PCI-VME Adapter for OpenVMS, for workstations and servers running
OpenVMS Alpha but have the PCI I/O bus. This device bridges Alpha environments to
the VME bus environment. Of special note is the XMI bus developed for the VAX 9000
and briefly discussed in the next section.
Because of its roots as an interactive multiprocessing system, the PDP-11 supported
a variety of operating systems. The RSX line (RSX-11M, RSX-11M-PLUS, and RSX-11s)
provided multiprocessing capabilities under priority-driven and/or real-time constraints.
In contrast, RT-11 was at the low end, offering multiprocessing in a single-user
environment.
While the PDP-11 maintained a respectable position at Digital until it was discontinued,
the favorite offspring is still the VAX. The first VAX offering was the VAX-11 line,
released in 1977. Whereas the PDP-11 was oriented more to the technical market, the
VAX was suited better for the commercial and business markets. The primary operating
system, VMS (Virtual Memory System), is a multiprocessing, interactive environment
well suited to multi-user business applications. The secondary operating system,
ULTRIX, is Digital's implementation of UNIX for the VAX. A real-time operating system,
VAXELN, is also available (primarily as a migration tool for PDP-11 users).
The VAX product line contains two segments: the Q-Bus MicroVAX line and the VAXBI
line. The MicroVAX line starts with the relatively old MicroVAX I and progresses
to the MicroVAX 2000 (which, unlike other MicroVAX systems, uses its own bus structure);
the MicroVAX II and the MicroVAX 3300, 3400, 3800, and MicroVAX 3900. These machines
are targeted to smaller businesses and departments and are also a bridge from PDP
to VAX technology.
With an eye to replacing the MicroVAX name and line, Digital introduced the VAX
4000 Series in 1990. These machines are positioned between the MicroVAX line and
the low end of the VAX line, and functioned as uniprocessing departmental servers
or high-end CAD workstations. From a performance perspective, the VAX 4000 Series
offers a price/performance ratio that is significantly better than the high end of
the MicroVAX line, while providing performance similar to the low end of the VAX
series. This overlap is part of Digital's strategy to reposition the low end of its
product line with its higher performance offerings.
In the big leagues, VAXBI systems start with the VAX 8200 Series, and then advance
into the 8300, 8500, 8600, 8700, 8800, and 8900 Series. In recent years, Digital
introduced the VAX 6200, 6300, and 6400 Series as the de facto replacements for the
low end of the 8000 Series (the 8200 and 8300). In fact, the 6000 Series was designed
to accommodate the multithreading of online transaction processing (OLTP). Additional
requirements for OLTP are addressed by the VAXft 3000, a fault-tolerant system based
on dual-processor MicroVAX 3000 technology.
At the top of the VAXBI line are the VAX 8600, 8700, 8800, and 8900 Series; the
9000 Series; and VAXclusters. While the high-end models of the older 8000 Series
approached mainframe capabilities, they had some difficulty maintaining fast access
to large capacities of memory and disk storage. These deficiencies were corrected
in the VAX 9000.
Most of the VAX series are upgradable to the newer Alpha AXP architecture, and
capable of running the OpenVMS or Digital UNIX operating systems. The Alpha AXP architecture
afforded a new 64-bit architecture that can directly address up to 1G of real memory.
The AlphaServer 2000 series (see Figure 2.2), based on Digital's new 64-bit RISC
technology, offers highly compact SMP servers with up to four processors; they are
used primarily as departmental or LAN servers. The 2000 series runs the DECchip 21064
CPU at 150MHz, and supports DEC OSF/1 AXP, OpenVMS AXP, and Windows NT. The 3000
Series, running the 21064 CPU at 175MHz, offer extensive memory and storage capabilities,
and is an excellent choice as a server for X Window terminals.
FIG. 2.2 Digital Equipment AlphaServer Family
The 64-bit design of the Alpha AXP architecture provides much more power to the
midrange, as well as added networking capacity and more file storage.
Top-End Offerings
Announced in 1989, the 64-bit VAX 9000 Series is Digital's best performing system,
offering mainframe-level performance. Using a combination of CISC and RISC technology,
the new processor and bus design (XMI) gave the 9000 a significant increase in data
processing speed. To achieve compatibility with the mainstream midrange machines,
the VAX 9000 also supports the VAXBI bus interface.
In addition to the main CISC/RISC system processor, the VAX 9000 supports specialized
vector processors for parallel operations. The Star Coupler CI interface was enhanced
for direct support of the XMI bus, and the resulting interface, termed CIXCD,
brings higher throughput and new configuration options to VAXclusters. In the most
general sense, the VAX 9000 was designed with one eye on compatibility between VMS
and the existing VAX line, while the other eye focused on optimization and mainframe-class
performance.
The RISC-based AlphaServer 8000 line has the ability to run up to 12 300-MHz 21164
Alpha processors, and combines the benefits of the mainframe with those of an open,
client/server system. It can accommodate a 10-terabyte cluster storage, and is usable
as a high-availability database server or OLTP server. This line includes the 8200,
which is scalable up to six 300 MHz processors, and can run both Digital UNIX and
OpenVMS. The 8200 and 8400 are the high end of the 8000 line, and are the first of
the AlphaServer 8000 series which Digital plans to extend through three generations
of products.
In designing the 8000 series, Digital had several goals, including support for
legacy I/O subsystems and DEC 7000/10000 AXP compatibility. The 8000 series doubles
the performance of the 7000/10000 AXP server line while providing a viable and fairly
straightforward upgrade path. In designing the 8000, Digital's design team conducted
a thorough analysis and performance study of the 7000/10000 systems. After the analysis,
Digital decided that system data bandwidth and memory read latency goals were critical
to their new design.
Read latency is the amount of time required for a CPU to read a piece of
data into a register in response to a load instruction. Cache memory is a
common way to minimize read latency. Latency tends to degrade as the number of processors
is increased, and latency will improve as the amount of available bandwidth is increased.
Comparable systems from HP and IBM do not stress low-memory latency in the design
of the RISC System/6000 or the PA-8000 SMP systems. Although the older 7000/10000
AXP systems were comparable to the RS/6000 and PA-8000 in terms of latency, the AlphaServer
8000 was developed with an emphasis on low-memory read latency. The 7000/1000s had
a latency of 560 nanoseconds, comparable to the ratings of the RS/6000 and PA-8000.
On the 8000 platform, however, Digital achieved a read latency of 200 nanoseconds.
In its design of the 8000 series, Digital pinpointed low latency as a major factor
in high system performance.
Symmetric Multiprocessing and VMSclustering
Previously known as VAXclustering, VMSclustering is executed with Digital's
OpenVMS Cluster Software. The software establishes an integrated OpenVMS environment,
which can be distributed over multiple Alpha and VAX CPUs. VMSclusters can include
both VAX and Alpha CPUs in the same cluster system.
To obtain high performance and greater computing capacity, Digital engineered
two optional processing configurations: symmetric multiprocessing (SMP) and
clustering. Both offer significant processing improvements.
SMP uses multiple processors sharing resources within a single VAX or Alpha AXP
system. The SMP architecture enables the system to break up functions and process
them independently. Thus, a disk access can be handled by one of the processors while
another performs a CPU-intensive calculation. The number of processors available
depends on the base model (note that not all models are capable of being used in
SMP mode).
Although SMP does provide some benefits based on its architecture alone, applications
must be specifically written (rewritten, or DEComposed as Digital calls it)
for the SMP environment. One final advantage of the SMP approach is that it makes
the system more fault-tolerant; if one processor malfunctions, the system can be
restarted and instructed to ignore the faulty processor. While this is far from an
online recovery operation, it sure beats being down until the hardware engineer arrives.
In contrast to SMP, clusters enable multiple systems to share disk storage in
a highly opti-mized manner. Clusters use a common piece of hardware, called a Star
Coupler, to provide a physically common point of contact between the multiple
VAXs or Alpha AXPs and the disk subsystem (see Figure 2.3). A VMScluster system can
contain an unlimited number of star couplers; the number of star couplers connected
to a CPU is limited only by the number of adapters provided by the CPU. A single
Star Coupler can interface multiple systems to one or more Hierarchical Storage Controllers
(HSC) which, in turn, control multiple physical disk drives. The interface between
the computers and the Star Coupler is referred to as the interconnect, a high-speed
link designed for fault tolerance. If there is an interconnect failure, the VMScluster
software will automatically resort to an alternate interconnect. The VMScluster software
will support any combination of these interconnects:
FIG. 2.3 Sample VAXcluster
- Computer Interconnect (CI)
- Digital Storage Systems Interconnect (DSSI)
- Small Computer Storage Interconnect (SCSI)
- Fiber Distributed Data Interface (FDDI)
- Ethernet
CI and DSSI are special-purpose interconnects, designed for CPUs and subsystems
in the VMSclusters. A maximum of 16 CPUs can be connected to a star coupler; a maximum
of four CPUs can be connected to a DSSI. The SCSI bus is not used for CPU-to-CPU
communications; therefore, CPUs connected to a multihost SCSI bus also require another
interconnect to provide this capability. The purpose of the SCSI bus is to provide
multihost access to SCSI storage devices.
Every aspect of a clustered system has built-in redundancy. And, the hardware-level
design can be enhanced further by doubling (or tripling in some cases) the number
of computers or amount of disk storage required. Overall, this makes for a highly
efficient, highly fault-tolerant hardware architecture. The cluster design provides
a self-contained backup system as well, and is an ideal solution to gradual expansion
because the cluster can be built one system at a time.
However, no degree of hardware tolerance can salvage the data damaged by an application
crashing or running amuck. And because the purpose of a cluster is to provide concurrent
access to a large pool of information, the integrity of that information is of paramount
importance--after all, there is no point in sharing a pool of garbage (unless you're
a goat).
To preserve the integrity of information, clusters can (and usually do) implement
the following safeguards:
- Resource locking. This allows various levels of locking on a clusterwide
basis. It allows file-level, record-level, and even field-level locking.
- Disaster rollback. After a disaster, this restores information to a previous
point at which it was known to be good. This feature usually works with recovery
processing, as described next.
- Recovery processing. After information has been restored to a known good
state, this feature reapplies the updates that occurred after that point and before
the disaster. For example, if an application inadvertently corrupted the information
at 12:05, the disk could be rolled back and then have the updates that occurred before
12:00 reapplied.
- Security. Maintaining an audit trail of what program applied what update
at what time is often a critical part of preventing a disaster from repeating. An
audit trail can find the offending program so it can be corrected.
On a much broader level, another technique to help maintain a high degree of data
integrity and availability is disk shadowing. With this technique, you attach two
disks to at least one (but preferably two) hierarchical storage controllers in a
cluster so identical information is maintained on both disk drives. Please note,
however, that you should use shadowing with the above recovery mechanisms to handle
situations in which the application corrupts both copies of the information.
As previously discussed, VAXclusters initially were developed to help Digital
compete in the higher end of the market. The advent of high-end systems from Digital,
however, will not make the cluster obsolete. Instead, systems like the Alpha 8000
Series will be used in clusters to maximize and even increase their high performance.
To sum up the salient difference between SMP and clusters, SMP is an architecture
that enables multiple applications within a computer to run in parallel (to a degree),
sharing common memory and disk resources. Clusters, on the other hand, enable multiple
applications on separate computers to share the same information stored on disk.
With clusters, there is no need to redesign or DECompose programs to take advantage
of the shared disk facility. Finally, SMP can be implemented within a cluster to
achieve maximum performance (provided the computers in the cluster support multiple
processors).
Strategy for Connectivity
Digital Equipment Corporation's strategy for connectivity is built upon DECnet.
DECnet is a set of products and services designed and implemented based on
the Digital Network Architecture (DNA). DECnet was originally implemented using proprietary
protocols, but in the early 1990s, Digital shifted DECnet into the open network environment
by incorporating support for both OSI protocols and for TCP/IP.
The shift toward open networking was the fifth major "phase" (revision)
in the DECnet specifications. The original DECnet phase (Phase I) was introduced
in 1976, and provided simple services for file transfer and program-to-program communication
between PDP-11 machines running the RSX-11M operating system. At the time, this was
fairly revolutionary, as software portability was almost non-existent. The subsequent
phases added more networking and interoperability services, bringing DECnet up to
the level of where it is today.
Phase II was released in 1978, and brought the capabilities of Phase I to all
major operating systems (including VMS). Additional features included remote file
access, network management tools, and support for point-to-point network configurations.
Phase III, released in 1980, introduced a dynamic adaptive routing algorithm, which
automatically calculated the best route to a message's destination. Under Phase II,
a direct connection had to be created between systems. Phase III also offered support
for X.25 packet switching networks as a method to connect systems, record-level access
over the network, and downline loading. A Phase III network could include up to 255
nodes.
Introduced in 1982, Phase IV was a major jump in technology. Under this phase,
Digital came to support ethernet, large LANs and WANs, virtual terminal services,
and various communications servers and gateways.
Digital's Phase IV approach to combining DNA with Ethernet (resulting in DECnet)
was the most revolutionary and sweeping networking architecture since IBM's Systems
Network Architecture (SNA). Ethernet, the result of a collaboration among Digital
Equipment, Intel Corporation, and Xerox Corporation, defines a physical interface
that supports a data rate of 10 million bps over a shielded coaxial cable. Up to
1,024 nodes (addressable attachments) are supported. One of the major benefits of
Phase IV was conformance with Open Systems Interconnect (OSI) standards.
The second important aspect of Ethernet is the collision-detection mechanism.
Because any node on the network could transmit at any time, a regulation mechanism
had to be created to recover from the inevitable collisions (which result in garbled
information). The technique Digital and Xerox put in place was termed the Carrier
Sense Multiple Access with Collision Detection (CSMA/CD).
Using the CSMA/CD technique, a node with information to transmit first listens
to the network. If no one else is transmitting, that node goes ahead and sends. If,
however, another node is transmitting on the network, that node waits for a predetermined
length of time before attempting to transmit again. If two nodes transmit simultaneously,
the collision is detected and both nodes wait for a random time period before attempting
to retransmit.
The CSMA/CD technique is also used in the IEEE 802.3 network implementation. In
fact, Ethernet and IEEE 802.3 are similar enough that nodes of both types can coexist
on the same physical network (although the two implementations are also different
enough that one type of node can't "understand" the other type's information).
Phase V was made available in 1991. It made significant changes to accommodate
TCP/IP as an alternative to OSI standards for open networking. Digital had three
goals in releasing Phase V: to permit a network to grow up to a million systems,
to incorporate both OSI and TCP/IP standard protocol suites to provide for a higher
level of system integration, and to support a distributed mode of operation. Although
a Phase IV network could accommodate up to 64,000 nodes, the industry's move towards
distributed, client/server computing soon limited Phase IV's viability. Phase V introduced
a new routing algorithm, which can potentially support millions of nodes. This algorithm
has since been adopted as a routing standard for both OSI and TCP/IP networks. Phase
V also set out to provide a distributed networkwide naming service, and to permit
nodes to generate their own addresses and register themselves automatically.
The Digital Distributed Name Service (DECdns) provides the following features:
distribution (so naming information does not have to be stored in a single location),
replication, dynamic updating, automatic updating, and a hierarchical naming structure.
Adding a new node to a Digital Equipment network is remarkably simple when using
DECnet with DECdns. The combination of these two products permit full autoconfiguration.
Where DECdns assigns names to computer systems, X.500 names extend to naming individuals
within a naming framework.
DECnet can be implemented on a LAN using Ethernet or over a WAN using a variety
of routers and/or gateways to bridge the distant LANs. Technicalities notwithstanding,
the primary function of DECnet is to deliver the following capabilities:
- Task-to-task communications. The capability of two programs (possibly
running on dissimilar systems or written in dissimilar languages) to exchange information.
- Remote file access. The capability to transfer files to and from remote
locations and to perform read and write (record-level) operations to a remote file.
- Network terminal access. The capability of terminals to access a remote
system and run programs on that system as if they were locally attached terminals.
- Network management. The capability to locate and isolate network problems
without bringing the whole network down.
- Downline loading. The capability to load a program or task from one system
onto another and run it on that system.
- Upline dumping. The capability of a system during abnormal termination
to send pertinent system information to an adjacent system. When the failed node
has been restored, this information can then be downloaded to help it resume its
normal operations.
Understanding the importance of DECnet is the key to understanding Digital's approach
to data processing. It is much rarer to find a single, isolated, non-networked VAX
than to find two or more Digital nodes communicating via ethernet.
Digital uses an X.500-based Directory Service, a general-purpose distributed directory
similar to Banyan's StreetTalk and Novell's NetWare Distributed Services. Digital
goes further than these two products in making the service accessible from any Web
browser or messaging system. Digital's service can be used to hold any information
that an organization wants to make public; sensitive data can be restricted. The
directory can be divided between multiple directory servers, with each server having
the ability to pass directory tasks to other servers that possess the relevant data.
Digital X.500 Directory Service supports Internet and X.400-based networks, and can
function as a superset of other directory services to form a global directory for
messaging and other applications. The service is integrated with Digital's MAILbus
400 backbone, which acts as a bridge between SMTP and other enterprise mail systems.
Developers can write applications to access the directory using the industry standard
X/Open XDS API. With the Digital X.500 Directory Synchronizer, directory information
can be exchanged and synchronized in a multivendor environment. Directories from
many popular e-mail products can be synchronized, and it can also support legacy
e-mail, Internet mail, and custom systems.
Messaging
Digital accommodates enterprisewide messaging in a multivendor environment with
a range of products that adhere to industry standards. Digital's messaging strategy
comprises a three-tier, client/server architecture, consisting of the enterprise
backbone, departmental system, and the desktop.
The messaging infrastructure lives at the Enterprise Backbone Server tier, and
provides mission-critical transport and directory services for heterogeneous mail
environments, as well as management of the messaging network. Connectivity to X.400
and the Internet is also achieved at this level.
At the Departmental Server tier, a mail or groupware server is deployed, and access
to the enterprise directory is achieved through a variety of servers, including MailWorks
(a client/server messaging system for LAN mail connectivity), Microsoft Exchange
Server, and ALL-IN-1. At the Desktop tier. E-mail clients can select from various
products such as Microsoft Mail, Lotus cc:Mail, TeamLinks, and Windows 95.
The Digital and Microsoft Alliance
Recently, Digital entered into a messaging alliance with Microsoft to develop a joint
enterprise mail infrastructure. The result of this alliance will be the eventual
integration of ALL-IN-1 and MailWorks, with the Microsoft Exchange Server through
MAILbus 400 and Digital X.500 Directory Service. MAILbus 400 is a store-and-forward
message transfer agent (MTA) that connects with other systems and services. MAILbus
is based on the X.400 standard, and can therefore connect to a wide variety of other
X.400 MTAs. DEC EDI (Electronic Data Interchange) provides connectivity between a
company and its customers or suppliers.
Digital and Microsoft's alliance further ensures the interoperability of all Digital
and Microsoft products. The Microsoft Exchange Server mail technology will be integrated
with Digital's ALL-IN-1 and MailWorks products, and Digital's mail backbone will
support Exchange Server. Digital will also support Microsoft's Windows Open Systems
Architecture (WOSA) API in OpenVMS, which will permit application developers to write
to both operating environments more easily. The alliance will contribute significantly
to the integration of Windows NT into the enterprise.
Application/User Relationship
Although Digital offers a variety of different operating systems and hardware
architectures, this section focuses on the most popular combination--the OpenVMS
operating system on Alpha AXP.
OpenVMS is an interactive operating system, originally developed for the
VAX, that gives each user the illusion that he or she is the only user on the computer.
Each user can run programs and access files independently of other users. If shared
services are required for access to a common data base, for example, the services
are handled at a system level, invisible to the user. After a user is logged onto
a given computer in the system, the further interaction between that user and the
computer is termed a session.
Within a session, the interaction between the user and the application is typically
character-oriented. Although Digital terminals and programs do support block-oriented
transmission formats for data entry, the application usually reads single (or small
groups of) characters from the keyboard as they are being typed.
For more sophisticated applications such as word processing, this reading method
enables the terminal and the application to interact as a PC keyboard interacts with
a PC--each key pressed can be interpreted as appropriate for that context. (This
similarity is one of the reasons that PC software vendors migrate toward DEC equipment.
Conceptually, they are very similar in intent and implementation.)
Although many other computer systems use a character-oriented, interactive interface
(for example, the HP 3000 under its MultiProgramming Executive, or MPE), Digital
has so refined this interface that it has become an integral part of Digital's applications
development strategy.
Digital has also developed its own graphical interface for terminals and PCs.
This interface, termed DECwindows, provides an alternative interface between
the user and the applications. As X Window graphic terminals become more available,
this method will become the preferred alternative to the traditional character-oriented
interface. DECwindows can be used with the OpenVMS, ULTRIX, and MS-DOS operating
systems.
Terminal Attachment Philosophy
Terminals can attach physically to the network through one of the following two
devices:
- A host. In this case, the terminal attaches directly to a Digital Equipment
host via an asynchronous link.
- A terminal server (DECserver). The terminal attaches (via an asynchronous
link) to a specialized device that manages the interface between the physical network
(ethernet) and the terminal.
The simplest connection consists of a terminal directly connected to a Digital
host via a simple, asynchronous line (see Figure 2.4). With this attachment, the
host responds directly to the terminal's activities without using any complicated
protocol. By necessity, the terminal must activate a session on the attached host
to access any local or remote application. The relationship between the terminal
and host is a simple one-to-one connection (one terminal connecting to one host port).
FIG. 2.4 Direct Attached Terminal
In the more normal case, however, terminals connect to terminal servers called
DECservers (see Figure 2.5). Although the interaction between the terminal
and its server is identical to the interaction between the terminal and a directly
connected host (specifically, no complicated protocol is required), the interactions
between the terminal server and the computers are quite different.
FIG. 2.5 Terminal Server Connection
Terminal servers talk to Digital host computers using the LAT protocol. LAT operates
independently from other DECnet protocols and provides two significant benefits:
- Group transmission. Under LAT, the terminal server collects characters
from a terminal and sends them to the host as a group, rather than as individual
transmissions. LAT transmission is often less disruptive for the computer hosting
the application and therefore can improve performance.
- Multiple hosts. Because LAT is not associated with any particular host,
the terminals attached to the terminal server can connect to any of the hosts on
the same ethernet LAN. In many cases, terminal servers provide a means for a single
terminal to invoke different sessions on multiple hosts and switch back and forth
between them.
More complicated are the cases in which a terminal directly connected to a host
wants to access another host, or when a terminal on a terminal server wants to access
a host that is not locally attached to the ethernet network but is attached to the
wide area Digital network. In both instances, another protocol, CTERM (Command
Terminal), comes into play. A DECnet host uses the CTERM to shuttle information
between a terminal and another host (see Figure 2.6). The drawback to this technique,
however, is that it consumes resources in the host sponsoring the remote link (that
is, the one initiating the connection).
Regardless of the type of connection, the terminal normally uses the XON/XOFF
mechanism to control the flow of data. With this technique, the terminal sends an
XOFF character to the host when it wants the host to stop transmitting and then sends
an XON character when it is ready for the host to resume transmitting.
Peer-to-Peer Relationships
DECnet, by its very design and intent, promotes peer-to-peer relationships among
its computing nodes. This peer-oriented relationship is at the core of the DNA that
underlies products like the VAXcluster and even DECnet; to obtain true distributed
processing, a network of peer processes (and therefore processors) must be established.
You can see how peer-to-peer relationships contribute to peer-to-peer processing
and communications by looking at remote file handling and task-to-task communications.
Regarding network file access, the entity requesting the remote file could be an
application program, the Network File Transfer (NFT) utility, or just a standard
copy command. In the case of an application program requesting a remote file, Digital
supplies a set of routines called the Network File Access Routines (NFARs)
to assist the program by performing some of the lower levels of the exchange. In
the case of user commands, the interface is handled via Record Management Services
(RMS).
FIG. 2.6 LAT and CTERM
On the other side of the equation (that is, the computing node with the file to
be accessed), resides a utility called the File Access Listener (FAL). FAL
listens for network requests for files on its node and translates the network request
into a local operation. FAL communicates with the requesting entity via the Data
Access Protocol (DAP), which is part of the DNA. In truth, DAP actually handles
most of the file transfer, another sign of how deep the peer-to-peer relationship
is embedded in DECnet.
The issue of task-to-task communications is, however, a bit more complex. A task
might have to communicate with another task that might not be running; or maybe it
is running, but is unable to respond. Therefore, programs performing this type of
network communications must follow some basic rules.
These rules, or task-to-task communications, begin with one program requesting
a logical link to another program and identifying the location (node) and name of
the target program. If the program is not known on the remote node or if the program
is not able to receive communications, the network will reject the link request.
If, on the other hand, the program is available and ready, then the link request
will be delivered to the program, and then the program will accept or reject the
request.
After the link has been requested and accepted, the programs can exchange data.
Information can be sent one or both ways--it is a function of the application, not
the network. When the communications are complete, one of the programs requests a
termination of the link, and that logical link is disassembled.
In 1990, Digital refined the marketing of its peer-to-peer relationships with
the introduction of Network Application Support (NAS). Targeted to compete against
IBM's Systems Application Architecture (SAA), NAS is offered as a set of software
products for developing and implementing distributed processing systems or client/server
applications. Behind the marketing hype is Digital's solid peer-to-peer technology.
PC Integration Strategy
Although Digital Equipment's PC product line has some shaky history, there is
nothing shaky about its PC integration strategy. Because Digital has such a strong
LAN foundation with its DECnet strategy, the integration of PCs is simply a matter
of interfacing them with the existing network. Furthermore, because of the wide presence
of Digital midrange computers in the general marketplace, PC LAN vendors such as
Novell have developed special products that enable Digital machines to participate
in their third-party network architectures.
Digital Equipment provides two approaches to integration (see Figure 2.7).
FIG. 2.7 PC-DECnet Integration with DECnet-DOS
and PCSA
To integrate PCs into the standard DECnet architecture, Digital provides DECnet-DOS,
which enables a PC to function as a DECnet computing node. DECnet-DOS offers the
following fundamental capabilities:
- Task-to-task communication. PC-based programs can use standard DECnet
task-to-task communications.
- Remote file access. The PC can initiate file exchanges with other DECnet
computing nodes using the NFT utility.
- VT220 terminal emulation. The PC can act as a terminal to access DECnet
applications using the SETHOST utility. This utility makes the PC appear as if it
were a terminal, physically attached to the target system. SETHOST also provides
full local printer support with Digital printers.
- Use of disk space. The PC can use disk space on a remote DECnet node as
if it were local.
A DECnet-DOS node can be connected to the network with an Ethernet or asynchronous
DECnet connection, using the PC's COM port as a physical link. The system to which
the PC is attached must be a DECnet Phase IV, full-function node, supporting asynchronous
DDCMP. For read-only operations, multiple DECnet-DOS nodes can simultaneously access
the same network disk.
The other approach is Digital's Pathworks, formerly Personal Computing Systems
Architecture (PCSA), which is a set of services that includes DECnet-DOS. Its benefits
(in addition to those already listed for DECnet-DOS) are as follows:
- Support for a variety of PC client workstations, including DOS, Windows, OS/2,
Macintosh, Windows NT, NetWare, LAN Manager, and Windows for Workgroups.
- Data sharing with UNIX and OpenVMS workstations, and VT or 3270 terminals.
- Support for a variety of servers, including UNIX, Windows NT, OpenVMS, and OS/2.
- Support for most network protocols, including TCP/IP, DECnet, IPX/SPX, NetBEUI,
AppleTalk, and LAT. Access to gateways is also provided for SNA and X.25 networks.
- The option to boot from the network so that PCs without hard disk drives can
be configured to load their operating system and programs from a VAX-based virtual
disk.
- Print server capabilities whereby the networked PCs can use the print resources
of the VAX server (in other words, the VAX running the Pathworks service software).
- File services that enable files to be shared between MS-DOS and VMS.
Both of Digital's approaches work for any PC-compatible device (in other words,
the solutions do not run only on Digital PCs). Also, both approaches require an Ethernet
interface card (from any of a variety of vendors) in the PC.
Pathworks permit PC users to run non-PC applications and access data beyond the
PC LAN. Pathworks' Network Connect interface permits PC users to access multivendor
file and print services. Further host connectivity is offered with services that
include terminal emulation, an X Window System server, and e-mail.
Several third-party PC LAN companies have been attracted to Digital's market.
This attraction, coupled with Digital's impressive installed base and the relative
ease with which software can be ported from the PC to Alpha, has spawned some interesting
products, such as Windows NT versions of OpenVMS and Polycenter Manager.
Office Automation
Digital's offering in this area is the VAX/OpenVMS resident product ALL-IN-1.
ALL-IN-1 is a menu-driven system that, in its most basic form, provides electronic
mail, calendar functions, and general information management functions. Furthermore,
ALL-IN-1 can be enhanced to include sophisticated word processing and spreadsheet
functions.
The basic package provides mail distribution and retrieval services to all participating
ALL-IN-1 users. The calendar functions provide for individual (or group, in some
cases) scheduling as well as general time-management functions. The information functions,
however, are quite sophisticated and include forms-management tools to make data
look attractive. Information entered into the facility can be used to create an online
reference facility, a public notice facility, or any other file-oriented application.
More often than not, ALL-IN-1's electronic mail function is tied to the advanced
word processing offered by the optional WPS-PLUS (usually pronounced "WIPS-PLUS").
WPS-PLUS is a word processor that uses the advanced capabilities of the standard
VT200-style keyboard to provide word processing features similar to those found on
dedicated word processors or PCs running high-end word processing software. These
capabilities include word wrap, cut and paste, find and replace, headers, footers,
multiple font types, and so on. To add even more sophistication, WPS-PLUS supports
optional modules for spell and grammar checking.
Another option to the basic ALL-IN-1 package is 20/20, a VAX-based spreadsheet
from Access Technology Inc. (Indianapolis, Indiana). Conceptually similar to the
well-known Lotus 1-2-3 spreadsheet, 20/20 offers a range of spreadsheet functions
and the capability to display business graphics. Some of the more advanced features
of 20/20 include project management (scheduling and tracking), as well as file interchange
with Lotus 1-2-3 (via a conversion utility).
Besides ALL-IN-1, Digital's Alpha machines come equipped with the following:
- POLYCENTER systems management and data center automation for networked systems
and clusters
- PATHWORKS for supporting Windows, DOS, OS/2, and Mac users
- LinkWorks for networking PCs, Macs, and Motif-based workstations and permitting
them to share data and applications
- ACCESSWORKS middleware for desktop access to distributed information
Network Architecture
Any DECnet LAN of reasonable size begins with a backbone. Individual connections
or tributaries of connections are dropped from the backbone, providing a relatively
understandable network topology that is simple to change or grow.
The backbone can be either a baseband (for computer data only) or a broadband
(for combined voice, video and data) medium. If a broadband medium is used, a
converter extracts the information from the broadband medium and converts it into
the more readily used baseband facility. The Ethernet baseband medium conforms to
the IEEE 10BASE5 standard.
Connections to the baseband cable are made through transceivers (see Figure 2.8).
Transceivers convert between the baseband medium and the ThinWire coaxial
medium and/or a 15-wire medium. The most common transceivers are the H4000 series
manufactured by Digital.
FIG. 2.8 Sample DECnet LAN
The cabling from the transceiver normally connects directly to an interface card
within a system or to another type of network device. The type of interface card
used depends on the bus of the host system. Some common interfaces that support connection
via transceiver and their matching buses include DELUA for Unibus (PDP/VAX) systems,
DELQA for Q-Bus (PDP/VAX) systems, DEBNA for VAXBI (VAX) systems, and DESVA for the
MicroVAX 2000.
Another common network device that connects to a transceiver is the Digital
Ethernet Local Network Interconnect (DELNI). DELNI connects up to eight devices
(via transceiver cables) to form a small LAN. DELNI can also connect to a baseband
Ethernet cable (via an H4000 transceiver) to turn the other connected devices into
tributaries of the backbone. (In smaller networks, DELNI is used alone or with other
DELNIs to form a LAN that has no backbone.)
To connect PCs, IBM PS/2s, or other devices that operate in locations (or in numbers)
that make standard transceiver connections difficult, Digital Equipment supports
the use of a ThinWire medium. ThinWire is a thin coaxial medium that conforms
to the IEEE 10BASE2 standard. ThinWire devices are linked on the coaxial cable via
a T-shaped connector for each attached device. Common thin-wire devices include the
following:
- Digital ThinWire Ethernet Multi-Port Repeater (DEMPR). The DEMPR provides
up to eight ThinWire segments. A DEMPR can be connected to a transceiver (or DELNI),
or it can be used by itself or with other DEMPRs to create a stand-alone, ThinWire
network.
- Digital ThinWire Ethernet Single-Port Repeater (DESPR). A DESPR connects
to a transceiver (or DELNI) on one side and provides a ThinWire segment on the other
side.
- Digital ThinWire Ethernet Station Adapter (DESTA). If a standard transceiver
device is present in the middle of a ThinWire network, the DESTA can translate the
ThinWire back into the standard transceiver interface. In a sense, the DESTA is the
opposite of a DESPR (which connects a ThinWire device to a transceiver).
- DEPCA. An Ethernet ThinWire adapter for a standard PC/XT or PC/AT (8-bit)
slot.
- DEMCA. An Ethernet ThinWire adapter for a PS/2 Micro Channel slot.
Terminals normally connect through terminal servers. Terminal servers can connect
to the backbone via a transceiver or DELNI and interface to terminals via standard
RS-232C asynchronous connections that can operate up to 19.2Kbps. Almost all terminal
servers use the LAT protocol for terminals accessing computing nodes.
In addition to providing connectivity for terminals, terminal servers also allow
connectivity to serial printers and modems. Two Digital terminal servers are:
- The DECserver 200. Allows the connection of up to eight serial devices. Multiple
DECserver 200s can be implemented through multiple transceivers.
- The DECserver 550. A rack-mounted terminal server that supports a range of optional
cards to provide additional connectivity or special-case connectivity (for example,
connection of IBM 3270 terminals). A DECserver 550 connects via transceiver and can
support 16 -128 serial devices (through the use of optional cards).
Bridges, routers, and gateways can attach to Ethernet via transceivers. These
highly specialized devices (normally computers in their own right) are used to implement
WANs or to provide interfaces with IBM equipment. Refer to Chapter 7, "LANs
and WANs," for more information on these devices.
Also, a Digital Data Communications Message Protocol (DDCMP) link can connect
Digital systems together over traditional telephone or serial links. DDCMP is an
integral part of DECnet that is often essential in implementing networks that are
low cost or non-PSDN (packet-switching public data networks). In a nutshell, DDCMP
is a system-to-system communications method.
Finally, because most of the components offered by Digital conform to the IEEE
802.3 standards as well as the Ethernet standards, a functional, mixed-vendor network
can be constructed with them. Such a network could use Digital Equipment systems
and networking devices but run alternative network software such as TCP/IP. (In fact,
because the equipment conforms to both standards, such alternative networking software
as TCP/IP and HP's NS can run concurrently with DECnet on the same physical network.)
Digital is the leader in Fiber Distributed Data Interface (FDDI) switching,
with its GigaSwitch product. FDDI is a high-speed fiber optic networking technology,
often used as a backbone for joining together servers or multiple LANs. GigaSwitch
is an FDDI switch that can significantly extend the useful lifetime of a LAN
by increasing the amount of effective bandwidth available.
Digital is embracing virtual LAN (VLAN) technology with its enVISN architecture
( for more information on virtual LANS, see the section "VLAN Technology"
in Chapter 11, "Network Management"). Under enVISN, policy agents
are created to ensure that the physical switch carries out quality of service, security,
and membership policies relative to the VLAN architecture. Quality of Service features
permit certain types of traffic to receive higher priority than others. For example,
a system with quality of service features might grant the highest priority to transaction
processing, and the lowest to file transfers.
© Copyright, Macmillan Computer Publishing. All
rights reserved.
|
