Continue Reading This Article
Enjoy this article as well as all of our content, including E-Guides, news, tips and more.
It's clear, industry hype around iSCSI is high; but what emerged as users talked it over was that it remains unclear just when -- and how -- they will actually be deploying it, particularly for users at the storage helm of midsized businesses.
Bill Slaughter, technical architect for Tupperware Corp., described his shop as squarely in the midrange -- big enough to have Fibre Channel (FC) networked storage, but small enough to be considering iSCSI for its cost benefits.
But, Slaughter said, while he was interested in iSCSI in theory, he remained hesitant about how to put it into practice.
"I already have an investment in Fibre Channel," Slaughter said. Tupperware has a FC SAN made up of two EMC Corp. Clariion CX 600s attached mostly to Linux servers. Slaughter said he had become even more intrigued by iSCSI now that more and more servers are coming packaged with iSCSI cards.
"But still," he mused, "I already have this infrastructure in place. I can't just rip it out and replace it with iSCSI -- that would defeat the cost benefits of putting in an IP SAN."
Slaughter said it was possible that when he expands his storage, he could make it iSCSI rather than FC, but he said his company's storage demands are only increasing 10% a year. It would take quite a while for that rate of growth to justify a new SAN, he said.
"It's not really about growth," he said. "Where I think it's most likely I'll consider iSCSI is when it comes time to decide whether or not to upgrade to 4 Gbps Fibre Channel.
"But I don't know," he said. "It's a tough decision. We're already using proven [FC] platforms."
Slaughter is not alone, at least among his cohorts at SNW. While most users gave an enthusiastic "yes" when asked if they were interested in learning about iSCSI, most also were at a loss when asked exactly how and when they would deploy it in their environment.
"We're going to need more storage for DR [disaster recovery]," said Mahesh Venkata, systems engineer for the University of Pittsburgh. "But we have some high-end, mission-critical applications I'd be nervous to put on iSCSI. I don't really see the use for it yet -- I have to learn more."
Analysts attempted to dispel the confusion with some insights into the iSCSI market.
"The very high end and the very low end will be the earliest adopters," said Robert Gray, analyst with IDC. "Fortune 50 companies are already accounting for 50% of the market and using iSCSI for DR or Tier-2 apps."
Two users from big companies confirmed they already have iSCSI up and running -- Roscow Lin of Pfizer said his company recently implemented an IP SAN for DR; Jason Watson, a systems engineer for a major IT vendor, who asked that his employer not be identified, said his company had an IP SAN up and running for "Tier-2" applications. Watson called the technical process of implementing iSCSI "cake."
Another "green field" for iSCSI, Gray predicted, was the low end, where users don't already have FC SANs to rip out and replace and will be enticed by iSCSI's relative simplicity and low cost.
But even that is not so simple, at least not according to Roel Flora, superviser of the core server group at the University of California at Santa Cruz.
"ISCSI is cheaper than Fibre Channel, but it's still not cheap," he said.
Federated Insurance WAN/LAN administrator Patrick Taphorn said his company was considering iSCSI storage for remote offices, some of which only have two servers, and all of which are still using DAS.
"Technically, it's still cheaper to just buy a server with five attached disks."
"If it's just simplicity people are looking for, they should consider NAS," said Greg Schulz, founder and analyst with the StorageIO Group. "There is still some amount of complexity to implementing any SAN. Users need to think about whether or not they really need the block-level access, not just how it compares to Fibre Channel."
The DAS users also said they could at least envision a place for iSCSI.
"We like the flexibility of networked storage," Taphorn said. "We'd like to be able to cluster our remote office servers and use VMware on them -- using iSCSI can help us do that."
"We already use some NAS, but it doesn't scale as high as we want it to go," Flora said. "And the block-level access is necessary for the permissions and security measures we have to implement, and for storing database data."
So where does that leave the midrange?
"They will probably be laggards," Gray said. "They want something mature and will be slower to adopt than the very high and very low end."
Gray said most midsized users, like Slaughter, will be more open to implementing iSCSI as part of a larger project. And, he said, some benefits of iSCSI that get lost in the hype about cheap and easy could entice more users as adoption grows.
"FC continues to be delicate to configure," he said. "Early experiences with iSCSI so far show that the protocol is mature enough not to have to work with certified configuration lists. That could just be luck, of course, but if it holds true, it could prove very tempting."