Before |
After |
|||||||||||||||||||||
|
Date: August 18, 1993 To: Terry Garcia From: John Doe RE: Use of ThinNet Ethernet (10Base2) technology in XYZ labs
Summary Pursuant to our prior conversation, the material which follows is derived from an internal document that I have published to Company XXX. It discusses the use of ThinNet technology in project labs in order to meet the need for LAN technology in client projects which costs less while supporting more rapid and flexible deployment. This material was not written for general distribution, however it may be useful as a low cost way of answering many of the questions you are asked on this subject without our spending the resources to write something specifically for this purpose. This material is of a summary nature; it is not exhaustive, and it is not a technical study. It is just intended as an overview. If there is anything in this material which is not clear, or if there is anything else that you need on this subject, please let me know.
Discussion This material discusses the use of Ethernet ThinNet (10Base2) LAN technology instead of twisted pair (10BaseT) LAN technology for some test environments. Local LAN environments are generally characterized as follows: being physically local to one room; requiring a high degree of flexibility and rapid reconfigurability (including, in some cases, an unknown or rapidly changing number of LAN segments and LAN connections); being temporary (limited to the life of the project); needing rapid deployment (driven by short term over-optimistic project deadlines); and having low budgetary support (needs to be cheap, since it wasn't originally planned for). ThinNet is preferable to twisted pair in these environments because of the following characteristics; it can be laid down readily in a single room environment (going beyond one room can be a challenge - depending on the physical situation) in come cases it will be easier to extend from one room (or building) to the next by bridging to twisted pair and jumping up to Internet, then back down to ThinNet in the next room); it can be put together, segmented, and reconfigured quickly and with ease (like tinker toys), and the number of local LAN connections it can support is practically unlimited for normal lab requirements; it can be knocked down, packed, moved, and stored with great ease for the next project; it is quick to configure and deploy by relatively inexperienced personnel (basically, any junior engineer can configure and deploy it in a matter of minutes); it is extremely inexpensive to buy for single room environments (typically, $10 - $2- per piece, plus $1 - $3 for connectors, many of which can be borrowed.) Twisted pair is better suited for more permanent installations, since it requires the use of twisted pair hub and/or wiring closet technology which ThinNet does not. The hub and wiring closet technology is generally more expensive (at least several thousand dollars per LAN segment) and is slower to deploy, since it generally involves the use of more experienced corporate resources that must be scheduled and are frequently unavailable in the time frame needed by the project. Twisted pair is also less flexible than ThinNet since most twisted pair hubs cannot be segmented and since a hub supports a limited number of LAN connections; both of these limitations can be overcome, but only with significant additional capital expenditures which have generally not been budgeted. In any case, since additional connections and segmenting for ThinNet is extremely inexpensive and since gaining the same capability for twisted pair is far more expensive and time consuming, ThinNet is the more cost and time effective choice of the two. Where connecting the local test LAN to Internet is needed, it can be achieved with a single Internet connection and a small 10Base2/10BaseT bridging transceiver which costs about $200, instead of having to bear the cost of providing a separate Internet connection for each device on the LAN. For example, in the case of the current XYZ project on which I am working, the requirement was to support somewhere between 24 and 28 LAN segments simultaneously in order to support simultaneous testing of multiple Ethernet connections for multiple database client/server configurations. There was a further requirement to dynamically merge and separate parts of these LAN segments on demand. Although XYZ was able to borrow a limited amount of twisted pair technology, it is infeasible to either borrow or buy 24 to 28 twisted pair hubs (potentially costing at least several thousand dollars each). By using twisted pair technology instead, XYZ has acquired a reusable asset for a cost that is less than that of a single hub. Together with the cost of the twisted pair cables, this has reduced project cost of LAN technology to less than 1/10th (perhaps 1/20th) of what is would be using twisted pair, for a savings to XXX Company of at least $50,000. |
Date: August 18, 1993 To: Terry Garcia From: John Doe RE: Advantages of ThinNet Ethernet (10Base2) Technology in XYZ Project Labs
Summary This document provides information you will probably find useful for answering many of the questions you are asked on this subject. Note, however, that it was not written as an exhaustive formal study for general distribution. If there is anything in this material that is not clear, or if you need more information, please let me know. LAN Environments
Comparison of ThinNet
vs. Twisted-Pair Technology
Table 1. Comparison of ThinNet and Twisted-Pair TechnologyConclusion Two of the twisted-pair limitations noted in Table 1the number of LAN connections supported and the ability to segment twisted-pair hubscan be overcome, but only with significant additional capital expenditures (which typically have not been budgeted for). Since segmenting and additional connections for ThinNet is very inexpensive, ThinNet is the more cost- and time-effective choice of the two. Where connecting the local test LAN to the Internet is needed, it can be achieved with a single Internet connection and a small 10Base2/10BaseT bridging transceiver, which costs about $200. This is less expensive than providing a separate Internet connection for each device on the LAN. For example, in the case of the current XYZ project on which I am working, the requirement was to support somewhere between 24 and 28 LAN segments at once in order to support the simultaneous testing of multiple Ethernet connections for multiple database client/server configurations. There was a further requirement to dynamically merge and separate parts of these LAN segments on demand. Although XYZ was able to borrow a limited amount of twisted pair technology, it is infeasible to either borrow or buy 24 to 28 twisted pair hubs (potentially costing at least several thousand dollars each). By using ThinNet technology instead, XYZ has acquired a reusable asset for a cost that is less than that of a single hub. Use of ThinNet technology has reduced the project cost for LAN technology to less than 1/10th (perhaps 1/20th) of what it would be using twisted-pair technology, for a savings to XXX Company of at least $50,000. |