EMC, NetApp, Sun to Manage Data Life Cycle
The H. Lee Moffitt Cancer Center & Research Institute is in the business of saving human lives, but in its data centers, its the life cycle of information that needs the ongoing care.
The Tampa, Fla., hospital, like many enterprises, mixes storage hardware and software technologies: 20 terabytes of StorageWorks XP 9900- and 9500-series disk arrays from Hewlett-Packard Co., content management tools from EMC Corp.s Legato Software division and from Storage Technology Corp., Veritas Software Corp.s NetBackup software, and tape drives.
"We have little islands of this all around. Hopefully, one day well be able to tie it all together so it makes sense," said Dave Bratt, Moffitts technology architect.
Enter ILM (information lifecycle management), a concept that big vendors such as EMC, Network Appliance Inc. and Sun Microsystems Inc. are beginning to peddle as the next step beyond HSM (hierarchical storage management). ILM adds brains to traditional HSM in that it doesnt deal with just blocks of data but addresses management of the actual content of the data from its creation through its use and to its obsolescence, advocates say.
ILM is "a strategy for proactive management of information. It is not a piece of software; it is not an array. It is a process," said Mark Lewis, EMCs executive vice president of open software operations, at the Storage Networking World show here late last month.
EMC spent $3 billion in the past four months buying Legato and Documentum Inc. to build out its ILM strategy. The Hopkinton, Mass., company plans to combine the best features from the software of both companies to provide an enhanced management layer for its storage arrays, EMC officials said.
EMC is not alone. NetApp, of Sunnyvale, Calif., will announce its ILM plan next month as part of the launch of its DataOnTap 6.5 operating system, said Krish Padmanabhan, director of data protection and reference storage solutions. NetApps Fpolicy API, which currently has open, create and rename options for files, will likely be renamed ILM API and will add close, delete and rename features for directories, Padmanabhan said. An update to the operating system, due next June, will give NetApp support for the Network File System, plus an event notifications manager, Padmanabhan said. All this should improve developer productivity when using DataOnTap.
Also next year, NetApp will add a native file movement engine to the operating system but will continue to offer more powerful solutions through partners such as Veritas and NuView Inc., Padmanabhan said. NetApp will also offer ILM integration services.
For its part, Sun is one to two years away from commercializing software, code-named Honeycomb, that will add to its ILM story.
"Its designed to investigate how you control the behavior of a storage system from the applications side," said Balint Fleischer, chief technology officer of Suns Network Storage Solutions division, in Palo Alto, Calif. Honeycomb uses Java and focuses on vertical-market applications linked to Suns file system, Fleischer said. An early example of Honeycomb-related work is Suns Infinite Mailbox technology for e-mail archiving, he said.
"There [are] two aspects of [ILM]: What is it, and is there a need for it?" Fleischer said. "There are needs to treat data differently. ILM is a newer, more generalized version of HSM. What ILM has added is a richer set of policies that you can define."
Hitachi Ltd., which supplies high-end arrays to HP and Sun, made ILM-related news late last month. The Tokyo-based company unveiled co-development and resale deals with content management player Ixos Software AG, of Munich, Germany, and storage management specialist AppIQ Inc., of Burlington, Mass.
The need for enhanced data management tools was emphasized by a recent report that said 5 exabytes of new data were created worldwide last year, which translates to a 30 percent year-over-year growth since 1999. Some 92 percent of this data was stored on magnetic media, according to researchers at the University of California at Berkeley in this years "How Much Information?" study, which was funded by EMC, HP, Intel Corp. and Microsoft Corp. Like moving to a bigger house, as storage gets cheaper to acquire and manage, companies and governments will start archiving data that wasnt previously considered worth keeping, lead researcher Peter Lyman said.
While Moffitts Bratt experiences this buildup of data, he said he will be cautious with regard to vendors ILM claims. On one hand, "with some of our research here with bioinformatics, [data] is just going to exponentially grow," he said. On the other, he said, "the HSM/ILM thingthey may add features to it, but its all the same thing in my book. When vendors come in here and try to say fancy words, it doesnt really bother us."