What's old is new again

Sep 04, 2012 14:09

I'm engaged in a craven attempt to get an Apple TV. I think I might be able to hack it as a mythfrontend at some point... if I keep using MythTV. The following is a post about The Future of Converged Infrastructure. I've been wanting to talk about this for a while in any case.

In 1999 I worked for an ASP working on converged infrastructure. We used concepts from the past and combined them with new capabilities to give us scalability and control. We were sure that the market was ripe to offer computing services to other firms and even to individual consumers. We didn't call it SaaS or IaaS but it was certainly what we were offering with our automated server builds and software installations. We had monitoring, backup, dashboards, product management, the works. The missing piece was visualization and orchestration software and VMware had just come out with a snazzy product that we were experimenting with. We had also just received a $50k capital investment so the outlook was very good indeed.

Unfortunately the "investment" included words such as "Synergy", "Refocus", "Realignment" and "Opportunities". And just in time for the Internet bubble to pop, too.

10 years later, we're talking about Converged Infrastructure again, and now everyone wants in on it. And the technology is there to meet it too. Every device has a software API so they can be treated as discrete appliances and regarded in natural ways. If you can't issue a command to bring up many networked machines and storage in a keystroke, you are not going to make it in the market. Now it is becoming common to not only bring up the infrastructure but to integrate your software as well and present it to customers all with a single command.

Some challenges are there for sure. Standardizing data protection, configuration management, patches and updates, real time monitoring and resource planning are still difficult to do. But with orchestration software if it can be done once, it can be done thousands of times. Test and dev environments can be created at a keystroke and destroyed at a whim without having them languish unused for months. We can use the same foundational technology to "burst out" to other cloud resources with only the wallet as the limiting factor. With common tools we can avoid vendor lock-in and use whoever can provide the lowest cost at the time we are willing to pay.

Now that it can be done, customers will expect these miracles to happen again and again. What does that look like? We'll have to have rigorous change management, central configuration management, business review and agreement on priority resources. The hard problems will remain hard: managing expectations for initial configuration and maintenance upgrades, priority for data protection, getting the proper resources for testing, keeping the hype of the new from occluding the practices of the old.

Because of this, the need for technical prowess will only increase. The savings expected by business will be consumed with the need to scale ever larger and serve even more customers. Responsibility for resources will be an ever more contentious issue as companies alternately outsource and insource their infrastructures. We are at the dawn of a golden age for information. And it's going to be a long hard slog.

converged infrastructure, crochety old people, cloud

Previous post
Up