CHICAGO -- Information is proliferating in the data center at unsustainable rates, consultant and author Jon Toigo told attendees at this week's Storage Decisions conference during his keynote titled “Storage 2012: Time of the Infrastruggle.”
Toigo coined the term “storage infrastruggle” in a recent Storage magazine article to describe the battles administrators face in managing a modern storage infrastructure.
“I see the world in terms of a storage buzzsaw,” said Toigo, who is CEO of Toigo Partners International and chairman of the Data Management Institute.
He listed the following trends as impacting today’s storage infrastructures:
- Unmanaged data growth is wasting most of the capacity on every drive.
- Server virtualization has more than doubled storage capacity requirements.
- A “disk everywhere" dogma has resulted in half of disk capacity being used to replicate data from the other half.
- Vendors are inflating the prices of commodity hardware.
- Vendor hype about storage clouds.
- Purchasing decisions based neither on facts nor strategic needs.
“Technology’s changing all the time; business priorities have shifted; ‘do more with less’ is the mantra -- they want you to reduce your budget by 15% while increasing service levels by 20%; and regulatory mandates impact how you store, encrypt and manage data,” Toigo said.
“You also have the issue of energy costs in some parts of the country where there isn’t enough of it," he said. "And there’s vendor woo -- especially around cloud technology -- and that’s increased our risk profile to an unacceptable level.”
Technology’s changing all the time; business priorities have shifted; ‘do more with less’ is the mantra -- they want you to reduce your budget by 15% while increasing service levels by 20%; and regulatory mandates impact how you store, encrypt and manage data.
Jon Toigo, Chairman of the Data Management Institute
Toigo said storage infrastructure management can be improved through proactive capacity growth and space reclamation strategies. “We need to stop just doing capacity consumption and start thinking about capacity reclamation,” he said.
While he lambasted hyped technologies such as cloud (“no standards, hidden fees, unpredictable service levels”) and server virtualization, he recommended storage virtualization in the form of hardware/software-agnostic management, solid-state drives (SSDs) and the Linear Tape File System (LTFS) for easing storage infrastructure management.
He called for smarter storage pooling, with one pool optimized for speed and a larger pool for retention of rarely accessed data. SSDs drive the speed pool, adding performance and reducing power by providing more IOPS while consuming less power. And LTFS can act as tape NAS to enable more efficient management of retention data.
“Storage by itself will kill us,” Toigo said, pointing to market research listing storage as between 33% and 70% of IT hardware spending for large organizations. “We have to do something to blunt the growth of storage.
“Not all your data is mission-critical, not all your data needs to be available to you right after the first disk array fails," he explained. "A lot of data, perhaps as much as 80%, isn’t needed for 24 to 48 hours after the system dies. If you think about it in those terms, you realize replicating everything to disk doesn't make sense.”