Does anyone in SA do proper information lifecycle management (ILM)? Precious few, and that's unlikely to change. But with the Consumer Protection Act now law and the Protection of Personal Information (POPI) Bill coming sometime soon, companies will at least have to start pretending that they organise their information in tiers for legal and efficiency reasons.
Mike Rees, business development manager for Commvault, says attitudes are changing because of the legislation.
"An important thing to note is that data management is an IT problem, but information management is a business problem," he says. "The minute data touches an application or a person, the organisation needs to define who is going to manage it and how.
“How do organisations manage retention, and do it economically? Should it be on disk, in the cloud or on tape? I think organisations largely believe IT will take care of it. But compliance is a process and an attitude that needs to be company-wide."
Morgan Malyon, solution architect at Dimension Data, agrees.
"What's happened is that businesses haven't worried about it for so long because there's never been a need to. So they stick data wherever they can, whenever they can - until there's a necessity to bring that data back. So from a management perspective, it's never been a focus.
“But it's now grown to a point where it's a problem. Businesses have realised they've neglected it for so long that they can't afford to let the IT department keep running it. It's grown exponentially over the past five years."
The way Google was able to recover was from the last line of defence: tape.
Manfred Gramlich, Oracle
Manfred Gramlich, sales specialist at Oracle SA, sees ILM as part of a wider requirement.
"We're seeing a requirement for enterprise content management that incorporates information lifecycle management. Managing your data and the lifecycle of your data obviously brings in how you manage it across tiers of storage and so on. The main topic today is really ECM. What we're seeing is that organisations are starting to understand the value of content and data, but they still have disparate systems."
Although POPI and the Consumer Protection Act are prodding more business managers to take another look at their information strategy, it started earlier than that, says Hein Matthee, presales engineer at Attix5.
"A lot of this attitude change started with King III. Suddenly the directors became responsible for their actions and it wasn't just about the IT department anymore."
Dimension Data's Malyon agrees.
"Some big organisations have been caught out over the past couple of years doing things like price-fixing. When those scandals hit, it wasn't a case of the IT department being responsible, but the guys at the top of the tree. The IT department doesn't pay a fine, the company does. It's a business problem."
The danger then is panic buying, says Commvault's Rees.
"If it isn't handled correctly, organisations go into panic mode and go out and buy point products that will solve part of their problem, for instance e-mail archiving. But not all data resides in e-mail. A large part does but not all of it."
Hushmeera Rajah, technology specialist at Novell, has also seen legislation prodding customers to take another look at their data and how it's managed.
"For the companies or customers we speak to, it's never been an issue before. But they've become more aware of it because of things like the Acts and compliance issues. Additional unmanaged storage is adding costs to their data centres. And one of the things we've seen is that most departments don't understand what data they own. They don't understand the content that's there.
“Unfortunately, there have never been automated processes to manage the growth in data. We only have the tools now to see what kind of data is growing and who owns it."
Disk to disk
But as John Hope-Bailie, MD of Demand Data, points out, the problem of managing data going back years is not so simple.
"Trying to manage information that has been generated over many years through business processes is very difficult and of course people are likely to deploy new business processes," he says.
"What we're seeing is that a lot of the physical provisioning around storage is already quite sophisticated. When people are putting together a new process, perhaps as part of a service catalogue, they know what kinds of storage are available and they know where to allocate them, but it's still a physical provisioning.
“In time we may get to the kind of logical provisioning that relates to the Privacy Act and retention and classification. It's not the lingua franca yet but it's coming through. Maybe it's time for the operating system to force the classification issue."
For Oracle's Gramlich, there is a bottom line and it's deciding on the hierarchy.
"The bottom line is: if the company strategy is to keep everything, you need to decide how you're going to store it. Oracle's approach is you have to first look at classification. In a lot of organisations there's a new attitude, but in the past they used to throw money at storage.
“What's available: flash, high-performance disk, capacity disk and then tape. As data changes, you move it across tiers. You can't keep exabytes of data on spinning disk. Tape is still strong. Our strategy is disk to disk to tape. High-performance data, put it on flash; 30 days to 90 days, high-capacity disk. Anything over 90 days, put it on tape.
“Google is a good example. It uses disk to disk to tape. When Gmail was down, Google had data corruption across disks, which didn't affect tape. The way it was able to recover was from the last line of defence: tape."
The IT department doesn't pay a fine, the company does. It's a business problem.
Morgan Malyon, Dimension Data
Tape is still strong in the enterprise, agrees Dimension Data's Malyon.
"Tape has never been a bad technology. It's just been evangelised to be a bad one by vendors that sell disks. If it had been a bad technology, it would never have stayed around as long as it has, in whatever form.
“There's a need for it and in our environments, there will always be a requirement for tape. That's why every time you design a solution, you make sure there's a tape drive and enough cartridges."
Even Attix5 is looking at tape, according to Matthee.
"As a disk to disk vendor, we are also looking at a tape component as part of the solution because it's cost-effective for archiving. Obviously cloud throws a spanner in the works because you can't then back up to the cloud, but for offsite and archiving, it's fine."
Finding out what's important
Could enterprise ILM move into the cloud? Pieter Potgieter, presales manager at Software AG SA, says not completely.
"Most businesses in the medium to large enterprise tiers have lots of complex systems and they cannot afford to take everything to the cloud. For many of the workloads, you cannot take half a solution into the cloud. Yes, there are certain loads where it makes sense to go to the cloud but for most systems, unless all of them are there, the to-ing and fro-ing just doesn't work because of the bandwidth restrictions."
A common problem in enterprise storage is backing up data that doesn't need to be backed up. Dimension Data's Malyon says the first step in any storage strategy exercise is deciding on what's important.
"What we're trying to drive with our customers is not necessarily moving away from tiering or data classification but asking what applications for them are critical. What are their Tier-1 applications?
“So we ask them to rank them based on what they use them for, and how we tier their data depends on how they answer. It then helps us architect storage, archiving and backup. When people throw more disk at the problem, they eventually run out of space. What they should be doing is figuring out how to use what they have better."
Organisations go into panic mode.
Mike Rees, Commvault
And not all data is created equal, notes Gramlich.
"In large organisations, there are so many copies of data and content when all you really want is the latest version. Information Management research recently said that if bad data impacts on operations only 5% of the time, it adds a staggering 45% to the cost of operations. We tend to view content management as a function of IT when it's not.
“Not all data is equal: if the CFO sends an e-mail out or a spreadsheet, that should be aligned to a business process and, therefore, you will take decisions about who has access to it and there's a security discussion about it as well. Business is at least starting to take a bigger interest in this. A simple example: what is data warehousing? Essentially, it's a kind of content management and content mining but because it's important to the business, there's a big emphasis on it."
Commvault's Rees says it can be a straightforward approach.
"Your definition of whether a document is important or not and my definition will vary. But does the organisation have to keep it? The safe way is to say yes. So the organisation needs a system that will take control of the data irrespective of what it is and where it is and allow them to find it when you need it.
“A few years ago, deduplication technology was touted as the way enterprises could save money on storage.
As data was stored, the system checked whether it was unique or a copy of previously stored data. But dedup hasn't made much headway.
Hope-Bailie says it's too expensive: "One of the reasons deduping hasn't flourished as much as it might have, is that although the technology is great, I'm not sure that for the equivalent amount of storage you actually save money. A dedup solution in its own right has a higher cost than just raw storage. It generally requires a serious commitment upfront to invest in dedup."
Attix5's Matthee says the human factor must play the ultimate role.
"You can have the greatest technology and the greatest tools but ultimately a company has to have a policy in place so that those tools get used. Deduplication is not going to solve any problems with regards to storage; there needs to be a policy.
“We had a customer with terabytes of storage that they kept throwing more storage at. We asked them whether they really needed MP3s backed up. They decided not to and dropped their storage requirements by 50%. It sounds like a silly example but it's common. We provide tools but it's the company's decision to use them or not."
Share