Each year, AWS re:Invent is one of the biggest and best gatherings of public cloud computing professionals in the world.
For those of us on the ground this week, the event is akin to a week-long, back-to-school event where everyone leaves more knowledgeable than when they arrived—including “students" and instructors.
I connected with a lot of people who were talking about solutions designed to support the public cloud. For example, I just attended "Building Enterprise Cloud operations as-a-service with T-Mobile" where Miha Kralj, managing director, Accenture Technology Strategy and Arjun Shah, senior manager, Business and Technology Integration at Accenture explained deficiencies of the traditional IT operations model, and enterprise requirements needed to run mission-critical applications in the public cloud.
Because I frequently talk to clients about this subject, this session was music to my ears; it was exciting to be in the audience and see the crowd’s interest. Steve Hull, director, Cloud Strategy and Transformation at T-Mobile, joined them on stage to showcase T-Mobile’s successful cloud operating model and demonstrate how their "uncarrier moves" can handle high variable workloads.
Successful cloud adoption requires process and cultural change
Miha and Arjun explained that cloud environments are very different from managed hosting solutions or traditional on-premises managed environment. I could not agree more. Trying to turn public cloud into just another managed hosting environment or an extension of their data center in the public cloud is one of the biggest mistakes I see companies make. Miha and Arjun asserted that successful cloud operations requires a rather different set of skills, processes and technologies. If you don’t think making the transition to public cloud does not requires a change, you should re:Think.
Miha and Arjun said that in their experience, most large organizations face serious obstacles due to lack of operational and economic discipline of managing the cloud. I agree with them, as most companies don’t realize they have an issue until 10 to 12 months after they have started leveraging the public clouds at scale.
The important thing is not how often issues arise, but rather that they are preventable. Companies just need to plan their cloud adoption holistically, from strategy and design all the way to operational excellence. This will make their journey to cloud smoother, and should not be a barrier to speed of adoption.
The real value in public cloud comes from agility
Users of public cloud want agile, self-service and elastic capabilities, not governance. As we know, a lack of governance can lead to serious issues. It’s the role of today’s operations team to make sure the public cloud is governed (not controlled), but with a new approach that meets current user requirements.
In addition to new methods of governance, successful IT organizations embrace an experimentation mindset, support multi-disciplinary teams and allow an agile creation of the modern cloud operating model that fits the needs and requirements of the environment. Service assembly, service delivery and service operations became the most critical capabilities that enable agile operations fulfillment for all tenants.
As Miha and Arjun explained in the session, modern cloud operations need to allow users to operate their assets with freedom and agility. That is the only way users will realize the full potential of AWS cloud, keep pace with technology innovation and avoid restrictive limitations of traditional operations. No one wants to wait six months for operations to add the services to a service catalog. They will find a work-around or a different path to get what they need.
The "white glove" treatment
Modern cloud operations should allow tenants to choose the best type of operational support. This can range from full, "white-glove" services for tenants with no DevOps aspirations, to minimalistic, self-service operations options, where tenants thrive in DevOps autonomy and independence.
A good cloud ops service catalog provides flexible choice of operations services, allowing each tenant to pick and choose the most efficient combination of monitoring, operating, optimizing and integrating services while still ensuring that the operation team can discover and analyze what is in use at any given time. They also need the ability to spending to prevent run-away costs before they become a substantial issue (think about managing your phone bills in the '90s).
To bring the whole story to life, T-Mobile’s Hull presented his company’s approach to cloud operations, which matches unpredictable workloads, such as device launches and retail events, while still meeting the operations requirements for a mission-critical application operated in an ITIL dominate IT organizations. By emulating this approach, enterprise operations can have its cake and eat it too.
This is a lesson all IT organization will have to learn. While it won’t be the only lesson this week, it is one of the most valuable nonetheless.
Find out more about our presence at AWS re:Invent.