Abstract

Full proceedings from the Tolly-sponsored 1994 industry conference. The conference was held in Boston, MA on June 14-15, 1994 and was chaired by Tolly Group founder, Kevin Tolly.

Topics:

High Speed Alternatives for LANs (IBM)

High-Speed LAN Developments and Trends (Thomas Conrad)

Remote LAN Access Technology (IBM)

Extending Client/Server into the Enterprise (DCA)

Virtual Connectivity for Branch Offices (Eicon Technology)

Bandwidth Optimization (ACC)

LAN Switching/Intranetworking Performance (SMC)

Intelligent Networks in the 1990's (Chipcom)

High-Performance Routing/Network Consolidation (Hypercom)

Integrating ATM into Enterprise Internetworks (Wellfleet)

Evolving Corporate Networks to ATM (CrossComm)

----------------------------

IBM — High Speed Alternatives for LANs
IBM’s first presentation compares the main high-speed LAN options available to enterprise planners in 1994: switched classical LANs, 100BaseT, 100VG-AnyLAN, and ATM. The presentation argues that there is no single universal answer; instead, the correct choice depends on the application and the part of the network being upgraded. IBM highlights switched Ethernet and Token Ring as practical near-term choices for collapsed backbones, server concentration, and smaller high-performance workgroups. It presents 100BaseT as attractive for preserving Ethernet framing and delivering 100Mbit/s desktop bandwidth, but notes limits in topology, management maturity, and asset reuse. 100VG-AnyLAN is positioned as a strong option for bursty, asymmetric traffic and environments needing priority handling, while ATM is described as the most scalable long-term approach for campus backbones, multimedia, and isochronous traffic, though with higher cost and migration complexity. IBM closes by tying technology choice to workload classes such as server centralization, high-bandwidth local applications, and high-performance peer-to-peer collaboration.

Thomas-Conrad — High-Speed LAN Developments and Trends
Thomas-Conrad’s presentation frames the high-speed LAN market as a transition from conventional shared LANs toward segmented, switched, and eventually multimedia-capable network designs. The opening material focuses on the need for high-speed LANs to the desktop, current market direction, and how technologies divide between backbone solutions and local/workgroup solutions. In the overview, Thomas-Conrad identifies ATM and FDDI as wide-area or backbone-oriented technologies, while 100Base-X, 100VG-AnyLAN, and 100Mbps TCNS are positioned as local-area or workgroup technologies. The structure suggests a migration model in which backbone and desktop needs diverge: some organizations need high-capacity aggregation, while others need more bandwidth directly at the user edge. The presentation appears intended as a market and architecture briefing rather than a narrow product pitch, helping attendees understand where each high-speed option fits in an enterprise roadmap.

IBM — Remote LAN Access Technology
IBM’s second presentation addresses remote LAN access and the design requirements for connecting remote and branch users into enterprise networks. The slides emphasize protocol flexibility and platform flexibility, stressing that a remote-access platform should connect to a LAN, connect to a WAN, support multiple protocols such as IP, IPX, NetBIOS, 802.2 LLC, and AppleTalk, and ideally avoid locking customers into a single server model. IBM also calls out special IP considerations, including network addressing, dial-in server behavior, per-user addressing versus pooled addresses, and the administrative burden of managing remote clients across different IP networks. The summary slide condenses the design criteria into protocol flexibility, OS flexibility, performance, security, ease of use, wiring closet implications, and management. Overall, IBM presents remote LAN access as an infrastructure problem requiring architectural flexibility rather than just modem connectivity.

DCA — Extending Client/Server into the Enterprise
DCA’s presentation focuses on how enterprises can extend client/server computing into wide-area and branch-office environments without abandoning existing SNA investments. The visible slides describe a “basic concept” and a remote node server strategy built around previously separate worlds: multi-host communications, distributed PCs, and the coexistence of client/server and SNA traffic. DCA’s message is that enterprise networking in the mid-1990s must bridge the old and the new rather than replace one with the other overnight. The material points toward migration strategies that preserve access to mainframe and transactional systems while enabling newer distributed applications and broader user access. The talk fits the conference theme of practical enterprise transition, showing how branch and departmental networks can evolve from traditional host-centric connectivity toward broader multi-service architectures.

Eicon Technology — Virtual Connectivity for Branch Offices
Eicon’s presentation examines the problem of virtual connectivity for branch offices where traditional SNA traffic and emerging client/server traffic must coexist across constrained WAN links. The slides explicitly state that SNA is today’s reality, that PCs and LANs will continue to proliferate, and that client/server computing is tomorrow’s reality. Eicon argues that these two worlds must be treated as separate but equal, and that “virtual connectivity” is the critical mechanism for integrating them. One slide, “Hierarchical SNA Meets Client/Server,” contrasts typical SNA leased-line traffic at around 9,600bps with client/server traffic that may require 64Kbps or more, highlighting a bandwidth mismatch at branch offices. Another slide discusses IBM’s APPN and AnyNet, but notes that they will have little practical impact before 1996. The conclusion is that enterprises must support both legacy host communications and newer LAN-based application traffic during a long transition period.

ACC — Bandwidth Optimization
ACC’s presentation is centered on bandwidth optimization as a practical WAN engineering discipline. The slides define bandwidth optimization as a way to minimize recurring bandwidth charges, maximize performance for a given WAN bandwidth, and mitigate network congestion. ACC treats bandwidth as the most costly network component and argues that optimization matters both for economics and for application performance. The deck addresses the congestion problem, presents express queuing as a mechanism for improving response for transactional traffic, and outlines multiple forms of compression, including header compression, data area compression, and full packet compression. It also identifies evaluation factors such as computational complexity, memory use, and latency, showing that compression is not automatically beneficial if processing overhead undermines real-world responsiveness. The overall theme is that WAN performance depends on traffic class awareness, queue discipline, and carefully chosen optimization techniques rather than brute-force bandwidth expansion alone.

SMC — LAN Switching/Intranetworking Performance
SMC’s presentation, “LAN Switching/Intranetworking Performance,” focuses on the role of LAN switching in improving enterprise network performance and segmentation. The title and surrounding slides indicate a discussion of how switching changes traffic patterns inside the intranetwork by reducing contention and creating more deterministic performance than shared-media LANs. The conference agenda places this talk in the section on newest advances in intelligent hub technology, suggesting that SMC framed LAN switching as both a performance tool and a migration step toward more intelligent wiring-closet infrastructure. In context with other talks at the conference, SMC’s emphasis appears to be on practical switching deployment, how it affects throughput and user experience, and how switched designs compare with traditional shared or bridged LAN environments. It serves as the conference’s main vendor section on switching as a campus and enterprise performance strategy.

Chipcom — Intelligent Networks in the 1990s
Chipcom’s presentation takes a broader architectural view, describing “intelligent networks” as the next stage of enterprise networking. The topic agenda calls out market trends and new technologies, and the visible slides cover subjects including SNA/LAN integration, high-availability environments, system architecture, and a plethora of technologies such as private dedicated Ethernet and ATM. Chipcom argues for a unified service-delivery platform that tightly couples end-node administration, the network operating system and server layer, LAN infrastructure, and WAN/internetworking. It also highlights high availability as a critical design goal, citing the need for fault tolerance and reduced outages. The slides suggest that enterprise networks should be built from adaptable building blocks capable of integrating new technologies over time, rather than around isolated point products. Chipcom’s message is that network intelligence increasingly means architectural integration, availability, and flexibility, not just port density or line speed.

Cabletron — Networking Infrastructure Using Intelligent Hubs
Cabletron’s presentation appears in the proceedings as a move from traditional connection-oriented LAN infrastructure toward more integrated and managed enterprise networking. The visible slides include “Redefining the Network,” “The New Information Transportation Infrastructure,” and “The Products,” along with material on stackable hubs. Cabletron positions intelligent hubs as a practical way to combine shared access, management visibility, and migration toward more advanced packet and cell switching. The stackable hub slide emphasizes branch-office suitability, integrated packet switching with bridge/router support, advanced management, and analysis through RMON, all with relatively straightforward deployment. The broader message is that wiring-closet infrastructure should become more than passive aggregation; it should provide the operational and management foundation for evolving enterprise internetworks. Cabletron therefore uses intelligent hubs as a bridge between simple shared LANs and more sophisticated switched or ATM-based network designs.

Hypercom — High-Performance Routing/Network Consolidation
Hypercom’s presentation addresses high-performance routing in the context of branch-office and enterprise consolidation. The visible title page identifies George Wallner, founder and chairman of Hypercom Network Systems, and a later slide titled “The Hybrid Router” summarizes the core message: there is no technological silver bullet, ATM will be too costly for many branch deployments in the 1994–1997 timeframe, and enterprises should instead combine several technologies. Hypercom advocates integrating Internet traffic, legacy protocols, voice, video, routing, controller functions, TDM, hubs, CSU/DSUs, and dial backup modems into a consolidated branch platform. This is a pragmatic consolidation argument rather than a pure speed pitch. The talk presents the branch router as a multifunction edge system that should absorb multiple roles and reduce device sprawl while preserving support for mixed traffic classes.

Wellfleet Communications — Integrating ATM into Enterprise Internetworks
Wellfleet’s presentation focuses on how ATM should be introduced into enterprise networks in a measured, hybrid fashion rather than as an immediate universal replacement. The visible slides discuss the evolution of the desktop environment, emerging network computing applications, hybrid environments with physically and logically defined LANs, and the expected penetration of ATM into enterprise internetworks. The presentation suggests that enterprise traffic is becoming more demanding because of collaborative computing, imaging, multimedia, and other bandwidth-intensive applications. Wellfleet positions ATM as an important future architecture for backbone and specialized high-performance environments, but one that must coexist with routers and switches for some time. The tone is integrative rather than revolutionary: ATM will matter, but adoption will occur gradually and must fit within mixed enterprise infrastructures.

CrossComm — TBA / ATM Migration Outlook
Although the agenda lists Larry Samberg’s session as “TBA,” the proceedings clearly include a CrossComm presentation and speaker biography. The visible slides show a strong message around enterprise concerns over response time, reliability, management, interoperability, and cost, and then build toward a conclusion that the telecommunications infrastructure will be based on ATM. Additional slides discuss after backbones, high-performance devices such as file servers and database servers, and the question of “ATM To The Desktop?” The final slide, “Slowly, But Steadily,” predicts a staged rollout: ATM campus backbones within roughly 18 to 36 months, ATM-based servers within 18 to 36 months, ATM WANs between facilities within 24 to 48 months, and broad desktop use only later. CrossComm’s contribution is therefore a forward-looking migration forecast: ATM is coming, but deployment will move outward from the network center rather than appear instantly at every desktop.