From a historical perspective, we started with clusters back in 1998. Clusters still had a fairly strong semblance of infrastructure within them. In 2006, that's when we did what you're all talking about here, which was to create that shared services organization and to pull all of the infrastructure out from the clusters into a central group.
I'd say that asset management is an ongoing challenge. You're big. You know that. You're very big. We are around 63,000 people. It's a hard enough job keeping track of what 63,000 people are using—and that's right down to a BlackBerry, a PC, a tablet, a server, or a mainframe. As you move into the data centre, you absolutely know what you have 100%. You know how old it is, how long it's been there, when it needs to be refreshed, and what version of software's running on it. You have to. If something goes wrong, you have to fix it. I think when it comes to our ability today, from an asset management perspective, I think we're in good shape. I'll never say it's perfect, but we're good. Prior to 2006, I would say that individual clusters, which had accountability for asset management back then, had a reasonably good handle on what they had. It was probably not perfect, but I think it was reasonable.
Asset management from a future perspective is something that we are always talking about and always challenged by, but that's far more on the software side than the hardware side. Software is where it's tough. Software worries me a lot more than hardware does, to be honest.