Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : McData (MCDT) -- Ignore unavailable to you. Want to Upgrade?

To: Douglas Nordgren who wrote (75)	9/21/2000 3:54:31 AM
From: Gus	Respond to of 234

I know. Your overall cost basis in Ancor was, what, <$1.00?<g> Regarding Inrange, they have a lot of legacy telecom and data networking switch product lines that they are winding down. Only 50% of revenues are related to storage networking, primarily ESCON director switches and channel extenders. INRG generated about $3.7 million in FC director switches during the June 2000 quarter. Like McDATA and CNT, the key to INRG's success will be getting designed into the SAN solutions of the leading RAID vendors. McDATA, of course, has the inside track in EMC and it is theirs to lose. Fortunately for CNT and INRG, Fibre Channel is an open standard and will have a much larger addressable market than the proprietary ESCON interconnect which was controlled by IBM. As part of EMC, McDATA clearly has the advantage of native interoperability with EMC's industry leading systems design which is characterized by high-availaibility and intelligence. EMC's Enginuity Operating System, for example, contains more than 1.2 million lines of code meticulously hardwired over the years in the the multiple high performance backplanes, the high-powered processor arrays, the cache memory, etc. Each of the 14 autonomous DataMovers in the Celerra file server contains a Pentium-based DART (Data Access Real Time) RTOS. McDATA's RTOS (Real Time Operating System) forms the foundation of its EFC Connectivity Server and Manager software. Those have to be designed to work and upgrade seamlessly with Enginuity and DART so EMC Software can essentially create a logical base map of the thousands of disk drives that will eventually be part of a large scale SAN. Then McDATA's highly integrated server/software/RTOS have to synchronize with the industry-leading statistical grid of EMC's failover and failure-detection system design in order to preserve the quality of EMC's industry-leading customer service which generally aims to catch the failure of components one day before it fails, at the very least. Here are some more details of about the way McDATA designed its RTOS: High Availability Real Time Operating System (RTOS) * Manages and controls failovers of redundant components * Proven ESCON heritage * 16 years of operation and support in data center environments * Field proven operation - real life feedback for fine-tuning and optimization - proven capability to operate in the data center where downtime is NOT an option All major field replaceable units (FRUs) have a redundant backup and are hot swappable - A failure in the primary FRU will not affect traffic through the director as backup component will automatically take over - two fully synchronized processors (CTP) - two synchronized memory modules with dual memory images (CMM) - primary and backup switch card (MPC) - Redundant power supply with loadshare and loadbalance (while both are in operation) - Redundant fans - dual power inputs Error detection and fault isolation - RTOS has the ability to track and correlate errors to pinpoint the exact location of a problem - error thresholds have been set based upon years of experience with RTOS in operation no frivolous failures reported eliminates transients as a source of troubleshooting headaches - inbound and outbound CRC checking isolates errors to sending device or Director hardware - Once a problem is identified corrective action is immediate automatic failover with switch card (MPC), memory module (CMM0 or processor (CTP). Director and and all ports remain in operation, data throughput not affected - bad port is identified and failed, no impact to remaining ports or data traffic Port cards are hot-swappable. If spare ports are available, little or no impact to the fabric otherwise impact limited to four ports only and for only a minute Troubleshooting is done for you by the director - no added expenses - customer support shows up and replaces affected component - no downtime Security and Data Integrity* Security is necessary for: +Heteregeneous operating systems on the same fabric +Business policy In the Director, both are addressed via Zoning + Segments the Director such that devices are allowed to only see the devices in their zone + Useful in preventing NT Servers from accessing and claiming all storage attached to the fabric +Prevents unauthorized access to other zones in the fabric Data Integrity in a switched environment is necessary to: +Prevent the right data from being picked up by the wrong port and vice-versa +The Director accomplishes this via separate, parity protected message paths between the switch card and outgoing ports. Message to outbound port to pick up incoming data expects acknowledgement, if any error, data is not allowed to leave the director users.nni.com This poses an interesting pickle for EMC's competition. Since EMC is generally considered to be the RAID vendor with the most reliable technology and the most flexible ability to add intelligence to its disk arrays, switches and software, can EMC's rivals really afford to use different switches? The 16 years of datacenter experience that McDATA has poured into its RTOS as well as its crossbar switch platform can't be easily replicated. Competing architectures like Brocade's large switch fabric and QLGC/Ancor's multi-stage switch can only hope to mitigate the intrinsically negative impact of ISLs (interswitch links) that have the practical effect of increasing the price per usable port as well as introducing unnecessary complexity into the failover and fault-isolation processes. There there is the question of circuit-level latency. Up and until the early August release of its own ASICs, McDATA had been using early generation ASICs technology --partly licensed from Brocade and partly derived from the HWP CNO acquisition. In its prospectus, McDATA indicated that the circuit-level latency of the ASICs they were currently using would impede the effective deployment of switched fabrics beyond 8,000 ports. Below is McDATA's transition away from its early-generation ASICs: 3Q2000 - 9-port FC switch 4Q2000 - 16- and 32- port FC switches 1Q2001 - 64-port director switch .....This second generation of FC SAN products provide the largest single stage switch in the industry.* In addition to the support for full duplex, 1.0625 Gbps ports at low latencies over Class 2 and 3 connections, the design supports 2.125 Gbps and future 10 Gbps links. This high performance non-blocking design is achieved through the use of Source Port Buffering, with Virtual Output Queuing (VOQ). This technique avoids head of queue blocking and port bandwidth issues. Message 14360332 McDATA, Inrange, Brocade, and CNT all use LSI Logic as their foundry so it's going to be interesting to see how the nascent fibre channel merchant ASICs market will eventually play out. Also, for reasons probably not having anything to do with incest, CNT, Inrange and McDATA don't seem to have the same problems making their director switches interoperate that OSFI -- McDATA, Brocade, Vixel, GAdzoox, Qlogic -- seems to have with, first, FC-SW-2 and now, with FC-GS-3.