Storage networkers tackle data issues

Some of the discussion you missed if you couldn't get to this month's Storage Networking World conference in Frankfurt.


After my comments about SNW-Europe the other week, it was interesting to see how most of the news this year was based around relatively low-level technology, while many of the presentation sessions were to do with relatively non-storage-specific topics such as ILM - the organisers had actually managed to find some real ILM users - and data management.

One of the keynote speakers, storage guru Jon Toigo (his blog is here), advised that users may need to hack into their storage arrays, arguing that they lack data visibility because they try to do clever stuff like virtualisation inside the box. He said that the in-box virtualisation can cause problems once you then virtualise again across the SAN, plus the lack of visibility into the data path restricts your ability to build storage systems optimised to the needs of your applications.

He was scathing of most current approaches to archiving too, describing them as the IT equivalent of a kitchen junk drawer, which you drop stuff into on the off-chance it'll be useful one day, but which you can never find anything in subsequently.

And he said that the next big thing will be the need for coherent deletion. The issue here is that once your data is old enough to escape SOX and suchlike regulations, you want to delete it - both to conform to data protection and privacy laws, and to avoid the possibility of 'smoking guns' being found by enemies using court orders to dig through your data. The challenge is that it has to be deleted thoroughly, from everywhere, even archives and off-site backups.

The other top speaker was Ian Angell, Professor of Information systems at the London School of Economics, who has hit the headlines in recent months with his trenchant criticism of the likely cost and unworkability of the UK national ID card scheme.

Among other topics, he discussed how technical solutions fix technical problems, not societal ones, that systems inevitably fail at some point, and that too many people mix up cause and effect, so they think that by measuring something they can product a real-world change - whereas in reality, people simply alter the model or change their accounting schemes to produce the desired measurement.

In a future column, I'll round up some of the new products shown at SNW-E.

"Recommended For You"

VMWorld 2008: VMware touts 'the cloud' at user conference Avoid VM stall, the scourge of virtualisation projects