Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

The Cloud Services Checklist: Automation, Service and Education

The CIO called a 4 p.m. meeting to gather his chief data architect, senior systems manager, applications group director, and QA and help desk managers over a fresh pot of coffee to talk about the finer points of cloud adoption.

This wasn't a meeting about "cloud in a box" or cloud services implementation. The real challenge now was ensuring that the company's cloud services strategy would pay off the way everyone believed it would. This meant understanding and being willing to do what it took to get there.

The first step was service.

Business users were already accessing public cloud services through portals for employee benefits and healthcare management. The troubleshooting process was simple: Call the cloud vendor and expect that service would be restored soon after the call. Users wanted the same service from the internal IT cloud. They didn't want to hear about problems and complications. They didn't want the finger-pointing, either.

Everyone sitting in that 4 p.m. meeting understood this "in their heads"--but the actual process of being part of a "service culture," rallying behind an application as a team and, in the process, forsaking the comfortable silos of technical expertise you had staked your reputations on--was an emotional and challenging issue. How would you be measured for professional excellence in a culture premised on service response, resolution and collaboration instead of on database technology know-how?

"Staff buy-in into the service culture was one of the areas I was most fearful of," says one financial services CIO. "In the first year's changeover to a 'service culture,' we lost three key contributors who simply didn't want to work under the new priorities and who had the market demand for their services that they could pretty much call their own shots and go elsewhere. And they did."

Bringing Automation and Control to Cloud Services

Cloud services also require the ability to quickly deploy applications. But often the only way to get there is by rattling the cages of staff and upsetting manual "best practices" for deployment that have been fine-tuned and diligently followed for years.

For example, the manual process that most shops use for creating a virtual image of a system is a hand-developed script. The script for deploying a virtual operating system like Linux modifies the OS to fit the destination platform and the application it will run. Developing these scripts is time-consuming. There's also the risk of introducing human error, since the process is manual. Beyond this, these virtual OS customizations pose compatibility risks with the original versions of operating systems that vendors commit to support.

Trials of new tools have drastically reduced the time it takes to both automate and virtualize deployment of virtual OSes. The automation also checks for compatibility with the underlying base operating system to ensure that vendor support is not lost. The catch, however, is that IT must get comfortable with virtual OS image creation through automation, and not through hand-crafted scripts that have been used successfully for years. This automation acceptance is not a "given," nor is it to be assumed in other areas of the cloud that are targeted for automation.

With cloud services, companies are infusing new automation to run workflows and established data center practices. One popular area of automation for financial services companies is in backup, failover and disaster recovery, due to stringent uptime requirements. They also increasingly look at automation to play a major role in auto-failover. Equens SpA, a European payments processor, wanted to move to as close to 100% systems uptime as possible by creating redundancy and failover mechanisms between its two data centers. It achieved this with automation that even delivered the ability for systems to detect failover situations and auto-failover without any management intervention.

"One day we might do this, but we aren't comfortable going quite that far today," says Guido Gatti, Equens SpA's general manager of operations. "We still want key managers to make the determination to fail over and to pull the switch."

Educating Users About Cloud Services and Pulling It All Together

IT managers may also be tempted to just let business users run their own clouds, since many public cloud contracts are subscribed to by business users without IT input. However, when the chips are down, it's IT that usually ends up with the responsibility of managing the cloud contracts and monitoring and troubleshooting service-level agreements and key performance indicators with the vendors.

"Once our business users get cloud, they turn it over to us to manage. This includes the vendor relationship. And if there is an outage, we get the call," says Carmen Suarez, director of the enterprise architecture services division for Miami-Dade County, Fla. On the flip side, most will agree that IT is best suited to handle technology vendor relations and results. Technology management is part of the IT pedigree. But the challenge now for IT in a cloud-based environment where business users want instant gratification is to make sure that the cloud investment is bringing acceptable returns, that the vendor is a good business partner, and that corporate governance, security and regulatory standards are upheld. Along with this goes IT-provided education to business users about the characteristics of great vendors so the best vendor selections can be made in the first place.

So why are these three areas--service, automation and education--so important now? Because all are critical to the health of the cloud. All are also relatively unappreciated by end users, who focus on reaping the benefits of cloud services and don't want to be bothered with technicalities.

For IT, this is nothing new. What is new are the pressures on IT professionals who are faced with a redefinition of which work is valued and rewarded; the redefinition (and automation) of work processes that are firmly established and that work well but are too slow; and the new emphasis on vendor management and other "soft" skills that IT must provide as part of the service culture. Few IT staff members are comfortable "in their own skins" at this point, but a modicum of discomfort always seems always to accomplish any major transformation, which cloud certainly is.