Operations Management in Converged Infrastructure

How do you manage operations in converged infrastructure?

There is a great deal of discussion about converged infrastructures and reference architectures and which vendors you should consider based on your needs.  However, if you still manage your single-SKU virtual solution the same as your legacy data center, using product or component-specific tools to run your operations, you are missing out on harnessing the power of this converged infrastructure building block—and getting the most out of your investment.

You need to manage your single-SKU as a single unit. Otherwise, you run the risk of having too many drivers going in too many directions and never reaching your destination—your cloud.

Fortunately, you can get a converged infrastructure management solution for a single-SKU designed for cloud computing. This management software automates and standardizes provisioning and operations across compute, network, and storage.

Too Many Drivers at the Wheel

Why is management for converged infrastructure important?

A single-SKU pre-integrated and pre-configured solution means tested and ensured interoperability of the components but surprisingly not integrated management as a rule. You are not going to get too far too quickly if you manage your single-SKU solution like a legacy data center, provisioning and operating network, storage and compute resources independently. You’ve got too many drivers at the wheel.

A recent IDC survey queried IT organizations about their top two challenges in moving to converged computing.  Not surprisingly, security and cost were high on the list but out of the top six results, four of the major concerns focused on management issues such as meeting service-level agreements (SLAs), downtime/availability, IT staff skill sets, and integrating with existing infrastructure.

If you step back and look at the concerns expressed by your peers, and then look at all the information around converged infrastructures, there is little focus on management software.  But, without the right management software, your new single-SKU is nothing more than an expensive sports car in need of a tune up.  Realizing the power and benefits of single-SKU solutions relies heavily on the right management software for monitoring, automation, orchestration, configuration, and other functions.

The highly popular VCE Vblock Infrastructure Platforms may best represent the next phase in the evolution of IT, one focused on simplifying IT operations while delivering more value to the business. However, to get the most out of these infrastructure platforms, you need management software that supports this evolution.

Getting a Good View of the Road Ahead

What’s needed?

While function-specific management software has its place in legacy data centers, single-SKU solutions beg for management software that harnesses the power of the complete environment to speed deployment, services, and ongoing management. Focusing on ongoing management where most of the activity lies, this software should give you a good view of the road ahead and provide at least three (3) fundamental capabilities:

  1. Monitoring: Operations management software needs to provide topological views across the heterogeneous platform environment, at any given time.
  2. Alerting: Centralized event management is critical for any cloud operations solution with integration into legacy enterprise management tools to provide a means to receive notifications about changes across the legacy and new cloud infrastructure.
  3. Analysis: Diagnostic capabilities are also required of an operations management solution to enable administrators to drill down and determine root-cause as needed.

Better Visibility

What’s available?

EMC Unified Infrastructure Manager (UIM) is one such tool. It is purpose-built for Vblock platforms to deliver converged infrastructure management for converged infrastructure. This EMC infrastructure management provides two core capabilities: 1. Unified Infrastructure Manager/Provisioning (UIM/P), and 2. Unified Infrastructure Manager/Operations (UIM/O).

UIM gives you better visibility with the ability to automatically provision, configure and change infrastructure services; manage compliance against VCE Vblock best practices; create topology views; determine root-cause and impact analysis; and integrate with enterprise management tools. We covered the basics of UIM, including the provisioning capabilities, in a recent post. But, standing up new services as covered in the first post is only a part of the management function, most of the day-to-day falls under operations.

UIM/O is already a multi-generation solution.  Initially, it provided:

  • Up-to-the-minute Topology Views:  UIM topology views provide visibility into your Vblock platform.  Current state views are critical in a converged agile infrastructure. UIM views are automatically updated as changes are made to the Vblock platform and services making it possible to quickly locate and understand the interconnections of Vblock components and services.
  • Centralized Event Management and Forwarding:  Integration with enterprise management solutions such as IBM Tivoli, HP OpenView, and BMC Remedy through SNMP forwarding of events and alerts bring end-to-end enterprise management to Vblock Infrastructure Platforms.
  • Automated Root-cause and Impact Analysis: UIM out-of-the-box rules identify root-cause within Vblock platforms. Event correlation is automated across all components to provide a single root cause, separating it from the symptoms and to show you what infrastructure service is impacted.

But, now with release 3.1, UIM/O offers much more.

Further On Down the Road

As pointed out in the previous post, while it’s questionable whether or not other vendors have yet to deliver even the basics of single-SKU management functionality, UIM/Operations is has already moved  down the road and is on to next-generation capabilities.

With the latest release UIM/O now offers:

  • Multiple Vblock Platform Management:  As multiple Vblocks become more common, IT staff can now monitor multiple Vblocks from a single console to further realize the benefits of improved operational efficiency.
  • vCloud Director Abstractions:  vCloud Director abstractions of virtual data centers synchronize with UIM/P infrastructure services to provide  topology views or maps. With the latest release, UIM/O augments alerts with vCloud Director abstractions to provide more information to operations for root-cause and impact analysis.
  • RESTful API Support:  Like the provisioning component, UIM/O now includes the EMC Data Access API (EDAA) interface to enable read-only access to UIM/O alert information to be integrated with other data center information sources into custom dashboards tailored to administrators’ needs.

Delivering Infrastructure-as-a-service (IaaS) architecture and realizing the promise of converged infrastructure means managing single-SKU Vblocks as a single pool of resources. Effective operations for these deployments require understanding the interconnections, visualization, monitoring, and automated root-cause and impact analysis.

On the Horizon

What’s coming in the future?

Once these critical fundamental management pieces are in place, monitoring performance, performing capacity planning and trending, and automating remediation will be the next critical steps in the converged infrastructure management evolution. Stay tuned, there’s more to come.

About the Author: Mark Prahl