Every once in a while (here, here), I have the same conversation enough times with customers that I find it useful to bring it to the blog community. Last week at VMworld 2011, I had the opportunity to spend quite a bit of time speaking with customers and partners about Intelligent Automation and what it means for IT to deliver self-service capabilities to their customers (internal, contractors, etc.). During the conversations, a number of questions came up over and over again:
Q - We're very distributed within IT and we'd like to get more centralized, more organized and more efficient at delivering services. Where do we start?
A - Let's assume for a second that you can work through infrastructure consolidation, consistency and virtualization - which isn't always simple because it probably crosses technology groups and organizational boundaries. Now you have a foundation that is ready to handle multiple types of applications and can be provisioned, expanded and de-provisioned without breaking the overall architecture. Now it's time to start thinking about "services", which means you need to think like a Product Manager. If a "user" is going to select a service, what do you want included in that service?
- Should they be simple services? (Order a Virtual Machine and Storage capacity)
- Should they be bundled services? (OS + VM + Network + Storage)
- Should they be tiered services? (Gold, Silver, Bronze pricing or service-assurance levels)
- Should they be metered services or unlimited? (time used, resources used, etc.)
- Should they be broad, complex services? (entire VDC, multi-tiered applications, PaaS environment)
This is where you begin thinking about the concepts that you would place into your "Service Catalog". The menu that you'll let your users select from. Eventually you'll get into other aspect of "Self-Service" (lifecycle management, capacity planning, etc.), but initially it's important to consider how a user-facing service can ultimately be packaged so they are more productive and IT can deliver it efficiently.
NOTE: This is where the term "IT as a Service" comes from. By leveraging the Stephen Covey concept of "Begin with the End in Mind", you're identifying where you want things to go before ever making a technology or operational commitment.
Q - Is "Self Service IT" an all-or-nothing proposition? What if we can't get certain groups or legacy applications to move to this new environment?
A - Unless you're working in a greenfield (all-new) environment, it's quite rare to find an organization that is willing/able to move to 100% "Self-Service IT" immediately, so the answer is that this can absolutely be done in phases or for portions of IT operations. Keep in mind that moving to this new model of operations involves both technology and people/organizational changes and change isn't always easy. This is where it's important to follow a few basic guidelines:
- Plan the initial steps/stages so that you can measure progress and quickly adjust as needed. This might be targeting a subset of services, or an early-adopter group.
- Once an early stage is successful, highlight the changes within that group that could be applied to other groups. During times of change, people tend to follow successes because the risk-levels are less.
- Reduce barriers to usage - for example, many groups struggle with how to bill for these services because they don't know what something costs internally today. Using a "show-back" model (identify costs, highlight them to groups or management) can be a first step that helps them better understand the value brought by IT and usage by their teams.
Q - What if my organization isn't in high-tech (no in-house developers)? Is there value that Self-Service can bring to my organization?
A - Absolutely. While many of the initial use-cases that are often highlighted for IT as a Service are for Dev/Test groups or basic IT infrastructure, almost anything can be made into a service.
- Internal IT groups are often the first users, as it helps them deliver services faster and with less mistakes, freeing them up to take on new projects or optimize existing IT services.
- We've seen customers create "on-board a new employee" services for managers, which includes ordering Laptop, Virtual Desktop (VDI), Unified Communications (UC) or Office Furniture.
- We've seen project managers leverage the systems to pre-allocate resources to align with an upcoming project schedule.
- We're seeing customers ask how it can help them simplify the onboarding and offboarding of contractors and associated resources.
The potential list of services that could be created (and automated) is huge, and it tends to grow as the capabilities are known within the company and evangelized by employees and managers.
Q - How much control can IT continue to maintain if everything moves to "Self Service"?
A - Since the Service Catalog and Automation/Orchestration system are under the domain of IT, the levels of control will be a function of how much value (above racking-stacking-cabling) IT delivers today and how much skill their customers have. Intelligent Automation is ultimately a framework for IT to have better business-level conversations with the business, where the answer is often delivered via technology. The system can be viewed as a way to offload repetitive tasks, or it can be an enabler for new business innovation. Either way, it's about IT delivering a competitive service within the business.
Q - Do these solutions ONLY work with Cisco equipment?
A - No, Cisco's Intelligent Automation tools are capable of working with many 3rd-party systems via APIs, scripts or CLI. While it works great with Cisco data center products like Cisco UCS and Cisco Nexus, it can also be deployed with 3rd-party storage, servers, virtualization, billing systems, etc.