ATM adapters give backbone to servers and nets -

 

Connecting state and local government leaders

A synchronous transfer mode adapters are great for the backbone, but they're not ready for the desktop. Deployed on a WAN, backbone or desktop, ATM is becoming the big gun of bandwidth. When National Software Testing Laboratories Inc. looked at the newest crop of PCI 155-megabit/sec ATM adapters, it found several useful server cards. But no vendor has yet mastered the formula for bringing ATM to the desktop. The prices of some adapters exceed the

A synchronous transfer mode adapters are great for the backbone, but they're not ready
for the desktop.


Deployed on a WAN, backbone or desktop, ATM is becoming the big gun of bandwidth.


When National Software Testing Laboratories Inc. looked at the newest crop of PCI
155-megabit/sec ATM adapters, it found several useful server cards. But no vendor has yet
mastered the formula for bringing ATM to the desktop. The prices of some adapters exceed
the cost of a desktop system.


ATM's golden promise is easy management of shared network bandwidth. Demanding imaging
applications could go over the same wire as low-bandwidth, latency-sensitive applications
such as voice.


That means agencies with separate Systems Network Architecture, voice, data and
Internet networks could combine them on one backbone. The network manager could decide
exactly how the available bandwidth would be consumed.


Fore Systems Inc.'s ForeRunner was our top pick for network management capability. The
ForeRunner followed the Microsoft Windows NT installation model very tightly. An emulated
LAN was added in the same way as a physical LAN adapter.


Fore's InForemation Center was the best utility for monitoring an ATM network in real
time. Because a good ATM analyzer can cost a year's salary, it's valuable to have this
free information about network health.


NSTL would like to see a number of improvements in the drivers--things that first-tier
Ethernet drivers provide.


Those things include easier remote configuration (which the lab did do in some fashion
on all the adapters through a bit of registry hacking), better built-in diagnostics, and
tighter integration with the Simple Network Management Protocol and NT's performance
counters.


The ForeRunner easily allowed the addition of virtual LAN adapters, either the ATM
Forum's LAN emulation (LANE) type or the classical IP-over-ATM type.


The drivers provided by Olicom Inc. made it easy to create richly functional
installations. However, Olicom's released driver package forces the LAN administrator to
create configuration files with a text editor. A publicly available beta utility puts a
friendlier shell around this text file creation. NSTL reviewers would like to see Olicom's
NT driver follow more NT configuration practices.


Most industry analysts have played taps for ATM to the desktop. There might be one last
hurrah in the Asynchronous Digital Subscriber Line marketplace, but it appears that ATM's
core market will be in backbones and public networks.


Using ATM as your networking backbone means accommodating a variety of workgroup LANs
and other types of traffic.


As a wide area networking unit, ATM, compared with the alternatives, is often the
cheapest way to get high performance and reliability. It also can be the best approach to
linking separate facilities with incompatible LANs over a public network.


Where available, switched service lets you control costs for bandwidth-hungry tasks
like videoconferencing by doling out the bandwidth on demand. This kind of switched ATM
service is rare in the public infrastructure.


If you have access to your network's fiber cabling, ATM can be a versatile means of
getting full use out of single-mode fiber.


WANs often have various Ethernet and token-ring flavors. All the adapters tested could
emulate both framing types.


Permanent virtual circuits are the rule rather than the exception in the public
infrastructure, so it's essential to have support for PVCs, preferably via
bandwidth-efficient classical IP.


Only Madge Networks Inc.'s software couldn't support classical IP on PVCs at the time
of the tests, but that likely will have changed by the time you read this.


You might want a card to route between a workgroup LANE and a WAN via classical IP.
When tested, only the ForeRunner and the Olicom RapidFire could handle this.


One driving factor behind ATM is the need for a high-speed, low-cost transport
mechanism for time-sensitive data streams of voice and video.


Instead of Ethernet, ATM competes with the circuit-switched architectures throughout
the world's phone systems.


A circuit-switching architecture ensures the lowest possible latency but deals poorly
with bursty data traffic. Circuit switching generally is more difficult and more expensive
to scale to gigabit speeds. Compared with time-division multiplexing circuit-switching
hardware, ATM is a real bargain.


T1 connections at a mere 1.5 megabits/sec can cost several thousand dollars each. T1
switches are considerably more expensive, port for port, than 155-megabit/sec ATM
switches.


That makes ATM two orders of magnitude cheaper than the circuit-switched alternative.


Unfortunately, critical ATM pieces are missing. No vendor had quality-of-service (QOS)
support, although Fore and Madge appear to be the furthest along. The reviewers assumed
all the vendors plan such support in the near future.


Fore had the most advanced support, though still in beta version, for the WinSock 2.2
specification. It provides constant bit rate support for the low-end ForeRunnerLE, but it
is unavailable for the high-end PCA card Fore sent for the tests.


Madge sent a beta driver with limited QOS support under NT 4.0.


Clearly, this a half-full, half-empty glass. Developers can begin working on
applications that take advantage of ATM without limiting customers to a single vendor. On
the other hand, users can't deploy these applications.


ATM gives the fastest available performance over copper wires. At 155 megabits/sec
compared with 100-megabit/sec 100Base-T, the best ATM adapters are 30 percent faster than
the best Ethernet adapters.


In situations where pulling fiber would be costly, this boosts performance compared
with that of 100Base-T.


ATM's CPU utilization can be higher or lower than conventional LAN architectures,
depending on how much processing support the adapter gives to the packet-to-cell
transition.


Forget bargain-basement client cards. Only the $1,000 ForeRunner could perform at wire
speeds in the tests. It had the best receiving efficiency at 19.5 megabytes/sec and the
second best sending efficiency at 37.4 megabytes/sec, averaged over all the tested
configurations.


The Olicom RapidFire was impressive whenever it was configured for large packets, but
it was unimpressive with standard Ethernet framing. Throughput and CPU efficiency for
Ethernet were half those for token-ring framing. The RapidFire achieved the best average
sending efficiency at 42.3 megabytes per CPU-second.


The Madge server adapter performed second only to the ForeRunner in raw throughput. It
had a voracious appetite for CPU resources, however. Madge adapters showed the lowest CPU
efficiency. The Madge client adapter managed just 10.7 megabytes sent per CPU-second
compared with the RapidFire's 42.3 megabytes sent per CPU-second.


The server card produced more reasonable scores of 14.7 megabytes received and 25.9
megabytes sent per CPU-second, a little worse than the adapters from Standard Microsystems
Corp. and Efficent Networks Inc., and significantly worse than the ForeRunner and the
RapidFire.


The Efficent and SMC adapters performed respectably over classical IP but miserably
over LANE.


The theoretical difference between classical IP and LANE doesn't explain this. It's
more likely to result from a driver difference that could be fixed relatively easily. Even
the classical IP scores were significantly slower than wire speed.


After extensive testing, the ForeRunner emerged as the ATM adapter of choice for
directly connecting servers and routers to an ATM backbone.


The Olicom RapidFire merited second place based on its reasonable cost, flexibility and
excellent classical IP performance.


The Madge Collage turned in impressive throughput numbers and should make your short
list, too.


No single card resurrected the idea of ATM to the desktop. Maybe ATM can ride directly
into homes on ADSL, but the business desktop still belongs to the various flavors of
Ethernet.


The cards that the reviewers liked are too expensive for widespread deployment, and the
one card designed for low cost didn't cut it on cost or performance.


NSTL's recommendations for ATM desktops are the same as for servers: ForeRunner,
RapidFire and Collage.


Here are individual comments about the tested products.


Efficient Networks Inc.of Dallas was a pioneer in PCI ATM adapters. Its products enjoy
widespread support in academia and good driver support in the fringe operating systems.


The company's adapter and driver proved rock-solid and covered all the basics in LANE
and classical IP.


With due deference to Efficient's achievements, this card is not one the review team
would pick. NSTL could not configure it to work with LANE and classical IP at the same
time. In fact, the reviewers had to reboot the test system twice in changing between the
two modes.


The classical IP implementation required configuring PVCs on the switch, which would be
inconvenient with a large backbone.


The Efficient adapter's performance left much to be desired. Its classical IP score was
respectable, but its LANE performance was among the worst. Classical IP CPU efficiency was
excellent. Ethernet LANE efficiency was poor.


Fore Systems Inc.of Warrendale, Pa., makes the ForeRunner adapter, which stood out as
the favorite. The speedy ForeRunner ran away with the performance tests. Its drivers were
among the easiest to configure, and it could handle up to 16 virtual LANs.


Fore's InForemation Center proved itself best. Like an always-available ATM analyzer,
it tracks the performance of the network connection in real time. None of the other
adapters offered comparable information.


The ForeRunner showed good manners in its use of the dual-CPU test machine. It spread
the load evenly across the two CPUs and did not interfere with operation of the mouse even
when fully loaded. CPU utilization was the lowest per megabyte of transferred data.


Taken together, this amounts to the ideal server card in manageability, flexibility,
high performance and low CPU utilization. NSTL recommends it for ATM server applications.
Of course, if you have the money, it makes a spiffy workstation card, too.


Fore sells a client card at an attractive price but did not send it for testing.
Perhaps this card, like the Madge client card, would come in at the bottom of the
performance charts. Given the cost of 155-megabit/sec switch ports, we recommend you don't
try to save a few hundred bucks on low-cost client cards.


Madge Networks Inc.of San Jose, Calif., sent a server card that could be one driver
release away from giving the ForeRunner serious competition. It had impressive LANE
performance and was among the easiest to configure. Madge's excellent documentation was
the most polished of any provided.


The test timing was unfortunate relative to Madge's development cycle. The test missed
a new release of Collage by two weeks. To its credit, Madge gave NSTL the go-ahead to test
with current drivers, which did not support classical IP and did not let reviewers set the
packet size for the LANE driver.


Despite the disadvantages, the Madge Collage took second place for overall performance.
Drivers were easy to install and configure. One spooky thing was their ability to use
relatively large 4,454-byte packets even when the switch, LAN emulation configuration
server, LAN emulation server and broadcast-and-unknown server were configured for
Ethernet-size 1,516-byte packets.


The Madge drivers demanded more system CPU resources than the other top performers.
Windows NT occasionally became destabilized under constant pounding from the Collage. The
mouse would often stop responding as the system struggled to cope with the barrage of
interrupts.


Clearly, alternate architectures will have to come into play as network adapters press
on toward gigabit/sec speeds.


The Collage client adapter boasted an ultra-integrated design that was little more than
a single chip and transceiver on a tiny PCI card. The ForeRunner and Madge server cards,
by comparison, contained complete reduced-instruction-set-computing processors on
generously sized PCI cards.


The Collage client's performance was equally diminutive. The reviewers don't see much
of a role for this card in clients or servers. Considering the cost of 155-megabit/sec
switch ports, NSTL recommends installing adapters capable of taking full advantage of that
bandwidth.


Olicom Inc. of Plano, Texas, makes the RapidFire, which NSTL recommends as a less
expensive alternative to the ForeRunner. The RapidFire drivers matched those of the
ForeRunner. Multiple virtual LANs and IP encapsulations could be configured to run
simultaneously.


Under Ethernet-framed LANE, the RapidFire ran out of the box without tweaking. It
performed competitively with the more expensive Madge Collage and ForeRunner. In CPU
efficiency, the RapidFire significantly outran the Collage and did almost as well as the
ForeRunner. It dominated the single-threaded tests, indicating that board and driver
designs were fine-tuned to reduce latency.


The Olicom card could not match the ForeRunner's raw throughput. It also performed
badly when configured for standard 1,516-byte Ethernet frames. And it did not always seem
to agree with the Fore Systems switch we used for testing.


Although the RapidFire driver supported switched access to
classical IP, reviewers couldn't configure this to work with the Fore switch and had to
use a permanent circuit.


The Olicom couldn't run when set for Ethernet framing with 4,454 bytes. The review team
reconfigured the card and switch for token-ring framing.


The RapidFire drivers were generally more difficult to work with. Olicom sent a beta
configuration program that improved ease of use dramatically, but no monitoring capability
comparable to Fore's was available.


Olicom officials said a new driver later this year will boost performance to equal that
of the ForeRunner. If so, that could put great pressure on Fore to reduce its prices.
Network managers will be even more impressed if Olicom can match Fore's InForemation
Center.


Olicom was the only company to ship LANE servers in the box--an interesting capability
if your switch doesn't already have integrated LANE support. Olicom's LANE server software
requires a dedicated MS-DOS PC.


Standard Microsystems Corp. of Hauppauge, N.Y., seems content to offer ATM as a
relatively small part of its broad array of networking gear. The Power155 ATM SMC earned
the same scores as the Efficient adapter because it is the same thing.


SMC's documentation was more polished, but the drivers were not current with drivers on
Efficient's World Wide Web site. NSTL recorded scores with Efficient's drivers, which were
faster and more reliable.


The Power155 adapter had the same strengths and limitations as the Efficient. Its
drivers supported classical IP, but not simultaneously with LANE. The system had to be
rebooted twice to change between LANE and classical IP, and performance was sub-par.


National Software Testing Laboratories Inc., an independent organization in
Conshohocken, Pa., provides information technology testing services to developers,
corporations and governments worldwide.


X
This website uses cookies to enhance user experience and to analyze performance and traffic on our website. We also share information about your use of our site with our social media, advertising and analytics partners. Learn More / Do Not Sell My Personal Information
Accept Cookies
X
Cookie Preferences Cookie List

Do Not Sell My Personal Information

When you visit our website, we store cookies on your browser to collect information. The information collected might relate to you, your preferences or your device, and is mostly used to make the site work as you expect it to and to provide a more personalized web experience. However, you can choose not to allow certain types of cookies, which may impact your experience of the site and the services we are able to offer. Click on the different category headings to find out more and change our default settings according to your preference. You cannot opt-out of our First Party Strictly Necessary Cookies as they are deployed in order to ensure the proper functioning of our website (such as prompting the cookie banner and remembering your settings, to log into your account, to redirect you when you log out, etc.). For more information about the First and Third Party Cookies used please follow this link.

Allow All Cookies

Manage Consent Preferences

Strictly Necessary Cookies - Always Active

We do not allow you to opt-out of our certain cookies, as they are necessary to ensure the proper functioning of our website (such as prompting our cookie banner and remembering your privacy choices) and/or to monitor site performance. These cookies are not used in a way that constitutes a “sale” of your data under the CCPA. You can set your browser to block or alert you about these cookies, but some parts of the site will not work as intended if you do so. You can usually find these settings in the Options or Preferences menu of your browser. Visit www.allaboutcookies.org to learn more.

Sale of Personal Data, Targeting & Social Media Cookies

Under the California Consumer Privacy Act, you have the right to opt-out of the sale of your personal information to third parties. These cookies collect information for analytics and to personalize your experience with targeted ads. You may exercise your right to opt out of the sale of personal information by using this toggle switch. If you opt out we will not be able to offer you personalised ads and will not hand over your personal information to any third parties. Additionally, you may contact our legal department for further clarification about your rights as a California consumer by using this Exercise My Rights link

If you have enabled privacy controls on your browser (such as a plugin), we have to take that as a valid request to opt-out. Therefore we would not be able to track your activity through the web. This may affect our ability to personalize ads according to your preferences.

Targeting cookies may be set through our site by our advertising partners. They may be used by those companies to build a profile of your interests and show you relevant adverts on other sites. They do not store directly personal information, but are based on uniquely identifying your browser and internet device. If you do not allow these cookies, you will experience less targeted advertising.

Social media cookies are set by a range of social media services that we have added to the site to enable you to share our content with your friends and networks. They are capable of tracking your browser across other sites and building up a profile of your interests. This may impact the content and messages you see on other websites you visit. If you do not allow these cookies you may not be able to use or see these sharing tools.

If you want to opt out of all of our lead reports and lists, please submit a privacy request at our Do Not Sell page.

Save Settings
Cookie Preferences Cookie List

Cookie List

A cookie is a small piece of data (text file) that a website – when visited by a user – asks your browser to store on your device in order to remember information about you, such as your language preference or login information. Those cookies are set by us and called first-party cookies. We also use third-party cookies – which are cookies from a domain different than the domain of the website you are visiting – for our advertising and marketing efforts. More specifically, we use cookies and other tracking technologies for the following purposes:

Strictly Necessary Cookies

We do not allow you to opt-out of our certain cookies, as they are necessary to ensure the proper functioning of our website (such as prompting our cookie banner and remembering your privacy choices) and/or to monitor site performance. These cookies are not used in a way that constitutes a “sale” of your data under the CCPA. You can set your browser to block or alert you about these cookies, but some parts of the site will not work as intended if you do so. You can usually find these settings in the Options or Preferences menu of your browser. Visit www.allaboutcookies.org to learn more.

Functional Cookies

We do not allow you to opt-out of our certain cookies, as they are necessary to ensure the proper functioning of our website (such as prompting our cookie banner and remembering your privacy choices) and/or to monitor site performance. These cookies are not used in a way that constitutes a “sale” of your data under the CCPA. You can set your browser to block or alert you about these cookies, but some parts of the site will not work as intended if you do so. You can usually find these settings in the Options or Preferences menu of your browser. Visit www.allaboutcookies.org to learn more.

Performance Cookies

We do not allow you to opt-out of our certain cookies, as they are necessary to ensure the proper functioning of our website (such as prompting our cookie banner and remembering your privacy choices) and/or to monitor site performance. These cookies are not used in a way that constitutes a “sale” of your data under the CCPA. You can set your browser to block or alert you about these cookies, but some parts of the site will not work as intended if you do so. You can usually find these settings in the Options or Preferences menu of your browser. Visit www.allaboutcookies.org to learn more.

Sale of Personal Data

We also use cookies to personalize your experience on our websites, including by determining the most relevant content and advertisements to show you, and to monitor site traffic and performance, so that we may improve our websites and your experience. You may opt out of our use of such cookies (and the associated “sale” of your Personal Information) by using this toggle switch. You will still see some advertising, regardless of your selection. Because we do not track you across different devices, browsers and GEMG properties, your selection will take effect only on this browser, this device and this website.

Social Media Cookies

We also use cookies to personalize your experience on our websites, including by determining the most relevant content and advertisements to show you, and to monitor site traffic and performance, so that we may improve our websites and your experience. You may opt out of our use of such cookies (and the associated “sale” of your Personal Information) by using this toggle switch. You will still see some advertising, regardless of your selection. Because we do not track you across different devices, browsers and GEMG properties, your selection will take effect only on this browser, this device and this website.

Targeting Cookies

We also use cookies to personalize your experience on our websites, including by determining the most relevant content and advertisements to show you, and to monitor site traffic and performance, so that we may improve our websites and your experience. You may opt out of our use of such cookies (and the associated “sale” of your Personal Information) by using this toggle switch. You will still see some advertising, regardless of your selection. Because we do not track you across different devices, browsers and GEMG properties, your selection will take effect only on this browser, this device and this website.