Information life-cycle management, intelligent storage-area networks and virtualization were some of the topics that dominated the thriving Storage Networking World conference in Orlando last week, which was hosted by sister publication Computerworld and the Storage Networking Industry Association.
The 2,800-plus crowd heard from the CEOs of prominent vendors such as Hitachi Data Systems and Brocade Communications, and users from a range of vertical markets, from healthcare to manufacturing and retail.
The presentations by these buyers served to put the whole storage puzzle in perspective. The problem crested in 2000 when data was growing out of control and companies were ill-equipped to deal with it.
The data store at Northrop Grumman, for example, had mushroomed from 5T to 20T bytes in four years and by 2000 was growing 75% year over year, says Paul Seay, chief infrastructure architect.
“The amount of data created compared to actual business growth was not proportional,” he says. “In the paper world we used to discard old information. We don’t do that well in the digital world.” So the company put the brakes on by getting smarter about what it stored and how it stored it, and today has slowed data growth to 35% per year.
That’s still significant, Seay says, but it’s a sign of the times. With more-sophisticated users relying more on smart applications, rich content and e-mail, he estimates each person generates 800M byte per year.
That’s nothing compared with requirements in the medical community. Robert Cecil, network director for the divisions of Radiology and Cardiology at the Cleveland Clinic Foundation, says he acquires 2T bytes of data per week. When mammography goes digital the average file size will be 100M bytes. He estimates that by next year the hospital will be adding 5T bytes per week.
Oh, and the data is never deleted. It is critical for the lifetime of patients and can be useful for long-term medical trend analysis.
All of which helps explain why companies are so interested in ILM, which is broadly defined as a way to classify information according to its business value, and uses policies to move it among tiered storage to optimize for performance and cost.
Seay from Northrop says ILM is 90% policy and 10% technology. “You need customers to understand the value of data so they don’t argue with you about how you handle it.” That’s why Northrop, in its ILM efforts, is focused on policy before technology.
