IBM keep trying..
IBM Looks For Storage Room In Government Corridors
(11/22/00, 5:32 p.m. ET) By Kim Renay Anderson, TechWeb News
In an effort to carve itself a bigger portion of the computer storage pie, IBM Corp. is chipping away at the decided market advantage held by its major competitor, EMC Corp.
IBM (stock: IBM) recently won a $500,000 contract to set up computer storage for the city of Boston, the latest contract it has signed with a government entity.
The city is using IBM's Shark Enterprise Storage Server to handle payroll and human resources records. Previously, IBM had signed public-sector contracts with the Food and Drug Administration and the Department of Defense.
"The public sector is important because it's not only a base of influential people, but you can have an impact on a great deal of people in this sector," said Clint Roswell, an IBM spokesman.
Jack Malinksy, Boston's director of operational technology, said IBM won out over a bid from EMC (stock: EMC) because Shark was immediately available. EMC couldn't deliver its system until early next year.
Analysts say that while IBM's public push is worthwhile, the company has a lot of ground to cover in challenging EMC. Years ago, IBM controlled the storage segment, but EMC has since overtaken it.
"If IBM is going to make headway into EMC's domain, it must grow its storage business as well as take customers," said Jack Scott, analyst at the Evaluator Group in Denver.
SMart move by Compaq I think....
Compaq Setting Up Home For SNIA
(11/22/00, 4:56 p.m. ET) TechWeb News
Compaq Computer Corp. (stock: CPQ) is building a home for the Storage Networking Industry Association. SNIA said Wednesday that construction has begun on its 14,000-square-foot building in Colorado Springs, Colo. Scheduled to open in January, the Storage Networking Industry Association Technology Center will showcase the development and testing of advanced network storage technologies that require interoperability of multi-vendor storage products. Currently in Mountain View, Calif., SNIA is a nonprofit organization of companies and individuals in the storage industry.
DAFS...
Section: REVIEWS
--------------------------------------------------------------------------------
DAFS: Just Daffy?
OLIVER RIST
Last year, SANs were the hottest thing going. Just last week, a local network designer scoffed that even thinking about NAS or SAN projects was folly, as both were dead issues. Years ago, I heard that the Internet was a fad, then I heard it would soon be dead due to too much traffic. And last year, e-commerce was a sure road to gold, while today they're all gasping their last breath.
I believe that if the idea is good, it'll survive-and network storage is not only a great idea, it's an evolutionary one. Without question, this is how internetworking needs to evolve.
There are only three basic problems that make network storage (both SAN and NAS) difficult right now. Most glaring is that it requires operating system support-anathema in heterogeneous networks. The proposed benefit of any network storage solution is that the storage acts as a generic network resource regardless of client platform or connectivity. Additionally, there is a performance requirement, where typical network read/write requests add significant overhead to the same requests issued locally. That means that high-speed transaction systems (a big target for network storage) usually want better throughput than an NAS/SAN implementation can supply. And, lastly, there are storage management schemes. Depending on performance requirements and underlying hardware, this chore can fall to such a granular level as to be heavily burdensome in both time and talent.
Do these difficulties mean SAN and NAS initiatives are phantom rainbows? Heck, no! You just need to wait until technologists catch up to the concept-and this past summer, it looks like they did. The acronym is DAFS, for the Direct Access File System protocol. DAFS is based around the Virtual Interface (VI) technology originally developed by Compaq, Intel and Microsoft to easily connect server clusters. What sets DAFS apart is that it addresses the performance issue of network storage as well as management issues.
By basing itself around VI, the DAFS protocol can handle direct memory-to-memory bulk data transfers that can bypass the overhead associated with typical network read/write requests. It will also be implemented as a file-access library to application developers, which means applications can be built from the ground up to be network storage-aware, essentially letting them send data transfers directly to the storage system regardless of any underlying OS requirements. And this isn't pie in the sky, either; the required SDK was made available by the DAFS Collaborative in late October at www.dafscollaborative.org.
Admittedly, DAFS is designed as a local area network storage protocol since the VI architecture requires an optimized local network like FE, FC or GbE. That makes it difficult to implement over the Web, although a clever front-end server load balancing and switching infrastructure should be able to compensate. Bottom line: DAFS may be the most forward-thinking enabling technology to network storage that I've seen to date. But even without it, I would never have dismissed network storage as a dead issue. Evolution takes time.
Oliver Rist is contributing technical editor at InternetWeek and vice president of product development at RCash In The Realm. He can be reached at orist@cmp.com. |
