Home > Blogs > VMware Operations Transformation Services > Tag Archives: VMware vRealize Automation

Tag Archives: VMware vRealize Automation

Key People, Process and Policy Considerations for vRealize Automation Success

Keng-Leong-Choong-cropBy Choong Keng Leong

Organizations implement VMware vRealize Automation (vRA) with the aim of shortening the provisioning of infrastructure services and the release of applications through self-service and automation. To achieve this, there is a need for balance between governance and business agility. Projects are more likely to fail or face significant obstacles if they do not plan adequately and ensure the necessary policies, processes and workflows are in place.

In this blog, we’ll explore some of these key planning and design activities that are often overlooked on the journey to cloud automation.

Key Players

vRealize Automation - Key PlayersThe very first thing we need to do is identify key players. The roles are mapped to actual team members in the organization. Minimally, we need to identify:

  • Service consumers – Authorized users of the self-service portal who can request and manage their cloud services, and which business groups they belong to
  • Approvers – Approves all possible requests
  • Cloud administrators Administers and manages the cloud infrastructure, cloud resources, and the configuration and maintenance of vRA
  • System administrators – Administers, configures and maintains the guest operating systems in the virtual machine
  • Application administrators – Installs, administers, configures and maintains the application software hosted on the virtual machine
  • Cloud security and compliance analyst –Monitors, analyzes and tests the security and compliance of application, guest OS and infrastructure

A common mistake is not identifying all the necessary key players and involving them in the planning and design early, which could have drastic impact to the vRA workflow designs.

Service Models

vRealize Automation - Service ModelsThe next step is to determine what cloud services will be offered through vRA. Many organizations start by offering Infrastructure-as-a-Service (IaaS), provisioning virtual machines leveraging existing vSphere virtual machine templates. For organizations that are heavily virtualized, this is not transformational and has very little incremental impact visible to the business.

To realize the full values of vRA, organizations should look beyond provisioning up to the OS level. The steps that follow after the server with OS is ready usually involve manual or scripted steps and multiple parties (app, middleware, db, security, etc.). Being able to automate these steps, package them and offer the package as a cloud service will result in significant efficiency gains. For example, instead of offering Windows 2012 as a catalog item, why not offer a SQL Server 2012 or a Tier 2 Application consisting of a pair of load-balanced Apache Tomcat Servers and a SQL Server?

Developing service models requires engaging the business to understand their requirements. For example, what is the point in offering a Windows Server 2003 R2 catalog item when no new business applications will be running on it. We also need to understand the service levels and performance requirements so that we can provision the machines in the correct pool of resources that provide these capabilities. We also need to identify which business groups will be entitled to these services.

Request Models

vRealize Automation - Request ModelsOnce the service models are defined, we can identify all the use cases for vRA and the types of requests within the scope of vRA. Request models (i.e. workflows) for the services are mapped out and documented. These may include:

  •  Request for a virtual machine
  • Request for a database server
  • Request to increase the resources of a virtual machine (e.g., add CPU, Memory)
  • Request to extend the lease of a virtual machine
  • Request to reboot a virtual machine
  • Request to decommission virtual machine
  • Request to snapshot a virtual machine
  • Request to back up a virtual machine

It is common to start by mapping out the current workflows and automating some of the steps using vRA and/or vRealize Orchestrator. While this approach may be quick, it has proven inadequate in many customer use cases I have encountered. Requirements to interface with a business system, process and function appear in late stages of the vRA implementation project, jeopardizing the project’s schedule and budget. In order for an organization to automate as much of the process as possible and make significant impact to service provisioning and delivery times, the whole service fulfillment cycle needs to be studied, optimized and transformed. It’s imperative to understand the whole business process through initiation of an IT/business project, budgeting, approval, procurement, installation, building, integration, testing, release, operation, management, support and retirement. Then, you must identify how the vRA will fit and interface with the various stakeholders, functions, processes and systems. Sometimes, it is necessary to have the vRA interface with external workflows already existing in other systems such as an IT service management (ITSM) system.

In addition, each request model needs to be correctly categorized and aligned with the organization’s governance policy and processes. For example, a request for a virtual machine in production vs. a machine for development will require different change management process, approval levels and approvers. These considerations should be incorporated into the design of the workflows and vRA approval policies. The request models can also be re-categorized to reduce governance overhead due to risk reduction with process automation and standardization of blueprints.

Access and Entitlement Management

vRealize Automation - Access & Entitlement ManagementAfter the key players, service models and request models are finalized, the different security access roles for vRA can be defined and mapped to the key players, so that they have adequate permissions and privileges to perform their tasks defined in the request models. Entitlements to the services are also configured and granted to the respective business groups and/or users.

Communication and Awareness

vRealize Automation - Communication & Transition SupportBefore the launch of the vRA, don’t forget to brief all key players on the processes and how to use the vRA based on their roles. Print and distribute reference cards and stickers to remind them of the process steps and how to get support when needed. It is important to cater for more hand-holding and support during the initial transition phase. The project will fail if users start to revert to old ways and stop using vRA.

========
Choong Keng Leong is an operations architect with VMware Professional Services and is based in Singapore. You can connect with him on LinkedIn

Transforming Operations and Perception of the IT Organization

By David Crane

dcrane-cropA recent engagement with a long-established telecommunications firm presented a huge challenge—the solution for which is a great example of how operations transformation can drive technical transformation. The firm’s customer base spans various global regions, each of which presented a different customer experience. The IT organization functioned in extremely siloed environments, having grown organically over 25 years to support an aging, fragmented infrastructure.

A frustrated but motivated CIO laid down the following requirements for the VMware consulting services team, to be met over an aggressive six-month timeline:

  • Reduce operational costs
  • Improve agility
  • Provide more service offerings
  • Help IT become a service broker and eliminate shadow IT
  • Build a flexible architecture to meet the needs of the business
  • Reduce total number of physical data centers
  • Gain more control and compliance of IT infrastructure environments

The internal IT team lacked the expertise and resources required to implement a software-defined data center (SDDC) solution. Their service request process was time-consuming, manual, and inconsistent. Add to that an average provisioning time for a full end-to-end server of eight weeks, and it’s no surprise that internal customers were seeking out external solution providers for their IT needs.

The VMware team set out to remedy all of this with the following solution:

  • Implement a production SDDC platform
  • Make self-service automated provisioning the first available service
  • Assess the customers’ operating processes
  • Introduce an optimized organizational structure
  • Integrate operations transformation and technical implementation
  • Take a phased approach to the project with clearly defined milestones to deliver immediate results
  • Ensure the VMware team team worked closely with internal groups

Transforming the Operating Model
Breaking down the siloed IT organization, and introducing horizontal, cross-departmental communications was the first step to allow the customer to become service-focused.

The team did have the business analyst concept, but the analysts sat outside the IT organization. They didn’t understand IT and weren’t incentivized to do so. As a result, rogue users were going out and doing things themselves, leading to compliance and governance issues.

We introduced the concept of infrastructure operations and tenant operations. These were cross-functional teams that talk to each other—a virtual center of excellence within the IT organization. As part of this organizational change, we brought in new roles, the two most important being the customer relationship manager and the service owner. We brought customer relationship management back into IT, so the person in the role started to understand IT and what they could deliver (and how) against customer requirements.

One of these requirements was the revelation that customers did not really have an interest in availability.  This was not because they didn’t care, but simply because IT over the years has become robust enough that availability is expected.  What their customers really cared about was the speed, and standardization, of the service provisioning lifecycle, as it was this that allowed them to quickly respond to market demands, and support the business objective to be the first to market with new products.

This led to a technical requirement as the IT organization’s customers requested to see this information in a dashboard format, so that proactive monitoring of the provisioning process could take place.

Transforming Infrastructure Operations
The service owners played a key role in saying VMware vRealize Operations only looks at infrastructure—this resulted in a demand to change things within VMware vRealize Automation.

However, the dashboards needed to be delivered through vRealize Operations. To meet the technical requirement, we focused on the self-service provisioning portal and allowed consumers to monitor the status of their ordered services via that portal. To do that, we needed a dashboard in VMware vRealize Operations to monitor the KPIs involved in service provisioning. In order to build the dashboard to monitor provisioning time, we had to create a custom solution using vRealize Automation. The technical solution was necessary to enable the operating framework architecture and organizational model to support it.

Dashboard Solution
We ended up with a provisioned resources dashboard as shown in figure 1 below that lists each virtual machine (VM) and the number of minutes it took to be provisioned. Less than 30 minutes shows green, less than two hours shows yellow, and over two hours is red. It also shows the average, minimum, and maximum times to provision.

Time to Provision

Figure 1:  Provisioned resources dashboard

The dashboard also enabled the customer to use data to feed back into the service life cycle process. For example, they started to understand service demand. Service owners—who were expected to forecast demand for services—could now do so with more accuracy. Now that the team was forecasting capacity demand more accurately, they were able to increase credibility by sharing this information with the infrastructure team. And ultimately they saved money by having a better handle on demand.

The dashboard also allowed IT to develop proactive operational processes.  On several occasions the service owners started to see a degradation in performance of the provisioning process, while the infrastructure monitoring dashboards were still showing a healthy ecosystem.

On further analysis, changes to the underlying infrastructure, whilst keeping in tolerance and SLA for the IT infrastructure teams, were having an accumulative impact further down the chain to the service provisioning process.

The provisioning dashboard and further integration with the customers’ service desk platform and event, incident, and problem management processes allowed the IT infrastructure teams to tune the change management process so that service provisioning would not be affected.

In the end, IT became service-oriented because of the dashboard. Because internal customers could use that tool to see the incredible accuracy with which the IT team was meeting its 30-minutes-or-less goal, it had a huge impact on the way the IT was perceived within business. IT’s credibility skyrocketed, and suddenly it became easier to drive things like the “cloud first” policy within the organization.

======
David Crane is an operations architect with the VMware Operations Transformation global practice and is based in the U.K.