Everything the cloud stands for is diametrically opposed to what enterprise computing stands for. But that’s okay — cloud promises to free up and expand enterprises in ways never imaginable with traditional systems.
How To: Buy a Pokemon Go Plus
Following the definition established by the National Institute of Standards and Technology (NIST), cloud computing “as an on-demand, elastic, metered and accessible resource computing service,” says Kemp.
However, from an enterprise computing perspective, “this is exactly the opposite of the way you think about computing,” he continues. “Enterprise computing is on-approval, static, private, purchased, and fundamentally inaccessible. Every single core tenant of cloud computing is literally opposed by a different philosophy in business. One of the challenges people have with cloud computing is it completely breaks every paradigm.”
Think about security, Kemp says. “When I think about security in an enterprise, I think about locking things down; I think about defining system boundaries; I think about understanding an isolating and putting things behind firewalls; I think about literally designing a security plan around a known system boundary.” When it comes to cloud, however, “I think about, ‘Wow, every Amazon server is available to everyone on the planet, even hackers in China and Russia.”
That’s not to say there is no common ground at all between cloud and enterprise computing. Ideally, as the business consumes services and resources, it really shouldn’t matter where those services are coming from — it may be the outside cloud, or it may be the corporation’s own internal data center. In either case, a solid architectural approach is needed to ensure that business requirements are being met, and those IT resources can be re-configured as those requirements change.
The increasing percentage of cloud in the mix calls for a complete rethink “about the way you build software and design software, and the way you design for reliability and security,” Kemp states.
For starters, any and all functionality and features is now based on software, not hardware. Hardware — especially servers — is something cheap and replaceable. “It would blow your mind if you knew how much a server costs Amazon or Google or Facebook,” says Kemp. “It’s not thousands of tens of thousands of dollars, it’s hundreds of dollars. When you have several hundred dollar servers, you literally treat them like cattle — if they fail, you leave them for dead. You don’t have service contracts, you don’t support this stuff.”
Instead, Kemp continues, “You design your software to tolerate failure. It’s a new paradigm in thinking where you have to literally take advantage of hundreds of thousands of machines that are all extremely inexpensive. Reliability and performance becomes something that you have to design into your software. You don’t rely on the infrastructure anymore to give you these characteristics.”
Another example of the opposites attracting within the cloudscape is devices. Cloud services are bringing together small, handheld devices and huge, back-end cloud data centers. These large data centers are, in many cases, “one big logical computer,” Kemp says. ”On the small end of the scale, were seeing pressure to have less expensive, more ubiquitous devices that surround us talk to ever larger services in the cloud. They enable each other.”
How To: Buy a Pokemon Go Plus