A
Andrew Piereder
Since this is addressed to me in a public forum, it merits a public response.
Mike Tennefoss,VP of Marketing at Echelon, apparently took exception to some earlier comments I made concerning Walt Boyes rather oracular pronouncements on the future of Ethernet at the I/O level. I had argued in part on the
issue of quality of service. Mr. Tennefoss seems to have a Cisco trademark and the generic technical sense of Quality of Service somewhat confused. Cisco addresses QoS for a very particular application, but it is unlikely
that their engineers would claim a high QoS for every possible networking application (including industrial and telcom).
A protocol designed to transmit video frames has a specific set of criteria that it must address to perform this function adequately. What would happen if a single video frame dropped out from some reason? Not much--the viewer would almost certain not notice a single drop out among thousands of frames. Hence a protocol designed to transmit video frames would not have a great
deal of engineering devoted to insuring error-free reception of each and every frame. On the other hand, and industrial protocol would find this an absolute necessity and would reflect engineering designed to accomplish this. Each protocol might be a masterpiece of engineering, but they would both be designed with a specific and perhaps mutually exclusive QoS issues.
Mr. Tennefoss also seems to confuse the adoption of LonWorks by companies that make industrial products, with LonWorks being used in industrial
applications (lets say MCCs...). Perhaps he is unaware of a possible motivation by some of these companies to leverage LonWorks compatibility to
exploit the HVAC marketplace? Does Wago really push LonWorks I/O interfaces into MCC applications for aluminum foundries or are the perhaps trying to make their product viable and attractive for a Building Automation project?
Adoption of specific fieldbuses by vendors is always a matter of their customer choices (or possible choices). It would be less disingenuous to list actual industrial applications where LonWorks has been implemented than
manufacturers.
Is Mr. Tennefoss claiming 25-35% of the industrial market? How is he defining industrial? No doubt LonWorks is a very successful protocol, but I think I am in a position to know whether LonWorks is a presence in
industrial markets. To be fair, and because I don't want to give the impression that I think unfavorably about LonWorks, I will mention that one of the finest applications of LonWorks technology is at Hach Company, a manufacturer of water quality testing materials and equipment. I worked closely with them last year to engineer a protocol bridge between their native LonWorks and the various common industrial networks employed by their customers. Hach use LonWorks as a backplane protocol as well as a way to network their various instruments (turbidimeters). They can cheaply and easily assembly instruments from various components simply by connecting them via LonWorks and employing a network manager to lay down the communication scheme. Hach basically engineered all tools themselves. Its a very impressive system and one well suited to the QoS issues Hach has--but its not a classic industrial application.
Clearly LonWorks meets QoS for a variety of applications, but its underlying technology and perhaps more specifically, Echelon's corporate culture, leave it at a disadvantage vis a vis other, more specifically targeted industrial
protocols such as ControlNet, Profibus, Interbus, DeviceNet and ASi.
Andy Piereder
Pinnacle IDC
Mike Tennefoss,VP of Marketing at Echelon, apparently took exception to some earlier comments I made concerning Walt Boyes rather oracular pronouncements on the future of Ethernet at the I/O level. I had argued in part on the
issue of quality of service. Mr. Tennefoss seems to have a Cisco trademark and the generic technical sense of Quality of Service somewhat confused. Cisco addresses QoS for a very particular application, but it is unlikely
that their engineers would claim a high QoS for every possible networking application (including industrial and telcom).
A protocol designed to transmit video frames has a specific set of criteria that it must address to perform this function adequately. What would happen if a single video frame dropped out from some reason? Not much--the viewer would almost certain not notice a single drop out among thousands of frames. Hence a protocol designed to transmit video frames would not have a great
deal of engineering devoted to insuring error-free reception of each and every frame. On the other hand, and industrial protocol would find this an absolute necessity and would reflect engineering designed to accomplish this. Each protocol might be a masterpiece of engineering, but they would both be designed with a specific and perhaps mutually exclusive QoS issues.
Mr. Tennefoss also seems to confuse the adoption of LonWorks by companies that make industrial products, with LonWorks being used in industrial
applications (lets say MCCs...). Perhaps he is unaware of a possible motivation by some of these companies to leverage LonWorks compatibility to
exploit the HVAC marketplace? Does Wago really push LonWorks I/O interfaces into MCC applications for aluminum foundries or are the perhaps trying to make their product viable and attractive for a Building Automation project?
Adoption of specific fieldbuses by vendors is always a matter of their customer choices (or possible choices). It would be less disingenuous to list actual industrial applications where LonWorks has been implemented than
manufacturers.
Is Mr. Tennefoss claiming 25-35% of the industrial market? How is he defining industrial? No doubt LonWorks is a very successful protocol, but I think I am in a position to know whether LonWorks is a presence in
industrial markets. To be fair, and because I don't want to give the impression that I think unfavorably about LonWorks, I will mention that one of the finest applications of LonWorks technology is at Hach Company, a manufacturer of water quality testing materials and equipment. I worked closely with them last year to engineer a protocol bridge between their native LonWorks and the various common industrial networks employed by their customers. Hach use LonWorks as a backplane protocol as well as a way to network their various instruments (turbidimeters). They can cheaply and easily assembly instruments from various components simply by connecting them via LonWorks and employing a network manager to lay down the communication scheme. Hach basically engineered all tools themselves. Its a very impressive system and one well suited to the QoS issues Hach has--but its not a classic industrial application.
Clearly LonWorks meets QoS for a variety of applications, but its underlying technology and perhaps more specifically, Echelon's corporate culture, leave it at a disadvantage vis a vis other, more specifically targeted industrial
protocols such as ControlNet, Profibus, Interbus, DeviceNet and ASi.
Andy Piereder
Pinnacle IDC