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 : Son of SAN - Storage Networking Technologies -- Ignore unavailable to you. Want to Upgrade?

To: David A. Lethe who wrote (3762)	8/6/2001 2:08:10 AM
From: Gus	Respond to of 4808

SPOF - single point of failure. Since you raised the topic, it would be more useful if you point out the SPOF in terms of the actual datapath: from the operating system I/O call to the SCSI/FC driver/adapter to the dynamic point-to-point connection created by the FC switch to the Channel Director to the Symmetrix Global Cache to the Disk Drive Director and vice versa. Otherwise, you'd just be as guilty of using marketing fluff in one of the most technically-oriented message boards in cyberspace. Note that the directors each connect to two independent memory regions on at least two memory boards while the data is moving through the global cache. ....all of the other features (except for the cache electonics) of their engines mentioned in your response are also available in even the $599.00 Mylex PCI RAID adapters. How 'bout the dedicated service processor running more than 1,000 diagnostic routines every 2 hours? Without the statistical information provide by those background operations, the level of automation that you find in the latest Symms would not be possible. Without that automation, I/O management would be a more laborious and tedious task requiring more $60,000-$150,000/year system administrators or more consultants like you. ....But for the record, IBM's shark doesn't have the common cache SPOF. Talk about marketing fluff! LOL. That's because Shark consists of two clustered (1+1 or N+1?) RS6000 computing engines inside one chassis. Each RS6000 uses shared memory for the L2 cache so your SPOF argument doesn't hold. There's also the question of using 2 servers to do storage's job, but I'll save that for the next time that an IBM supporter posts here. EMC's hardware is extraordinary. It isn't. Their software is. That's an antiquated distinction, David. The emerging consensus is that the scaling of networked computing is going to require better integration between hardware and software, not less. Even Microsoft and Intel understand this. Why do you think they're trying to shorten the well-known lag -- 2-3 years -- between Intel's processor cycles and Windows-based application cycles. ....If all of the clariion product was compatible with the symmetrix software, then clariion HW would probably outsell the EMC HW You're being simplistic and argumentative. 5 generations of Clariion were designed for Unix/NT/Novell environments. Clariion was never designed for the mainframe environment so why even bother? Symmetrix was originally designed for the mainframe environment and successfully ported to open systems as customers continued to migrate their secondary applications away from the mainframe in the 90s. Note that the S390 mainframe and Solaris are currently considered to be the most reliable operating platforms of most data centers today. Anyway, I don't mean to be contentious because the idea is to lay out the as much of the competing technology road maps beside each other. In this particular case, the flurry of informations is a result of EMC's analyst day. Since EMC is the industry leader, its technology road map will influence the plans of its rivals. Take this nugget, for example. .....EMC did not offer any details about upcoming products, but did give a glimpse of a new initiative aimed at automating storage. As part of its Automated Information Storage (Auto IS) initiative, EMC plans to add virtualization technology to its Symmetrix storage arrays, says Jim Rothnie, senior vice president and CTO. Also coming is technology to allow data to be migrated automatically from array to array, not merely from hard drive to hard drive, says Rothnie. Currently, the Symmetrix arrays have the capability to reallocate data from a drive which is heavily used to another less-busy drive, thereby rebalancing the data load and increasing performance, Rothnie says. However, new versions, expected to be unveiled by year-end, will allow new boxes to easily connect to current boxes, with data migrating from box to box automatically, either to smooth out imbalances or to quickly scale up in terms of capacity. To make this automated data migration happen, EMC will introduce new capabilities into its EMC ControlCenter software, and introduce new I/O redirectors to allow I/O to be automatically redirected as problem spots occur, says Rothnie..... crn.com In order to provide the same capability of automatically moving, EMC's rivals have to match EMC's breadth of storage operating systems from Enguinity in Symmetrix to Control Station/DART in Celerra to the Crosstor Operating System in Clariion because those three storage operating systems will have to work seamlessly with the fabric OS (e.g., McDATA's E/OS, Brocade's Fabric OS, etc) to allow those transfers.