Automating Infrastructure Management at EMC

What is storage-industry leader EMC doing to simplify and automate management in its data center?

EMC IT is deploying EMC, VMware, and partner management automation technologies to simplify operations in its data centers. This effort includes installing new products, adopting new processes, and changing roles to further evolve a private cloud model.

EMC practices what it preaches—and uses what it creates. EMC IT is leveraging EMC as well as third-party enterprise management, automation, and orchestration technologies to discover and manage performance across complex virtual server, network, and storage in a private cloud environment.

These infrastructure management technologies and the changes they bring to the global EMC IT structure would be unwieldy if not directed by sound principles. An internal-customer orientation with better services as a goal guides the EMC IT changes to the many areas that comprise its infrastructure management environment.  An internal governance team makes technology decisions with the entire IT organization in mind. This approach sets a good foundation for the present and for adopting new technologies in the future.

Customer First

EMC IT has fully embraced the cloud model as demonstrated in day-to-day activities. From internal portals to mobile apps for employees and customers, this IT department is not the same one I observed when I joined EMC 5-years ago.  Back then, mobile apps were for consumers only, portals were still mostly found on the Internet, and Apple devices were used by only a few insistent senior managers.

Back Office Makeover

What’s changed?

The move from silos to services involves a lot of moving parts. Data centers are no longer limited to specific servers running specific applications tied to specific storage. Cloud environments abstract physical hardware from software where much of the data center intelligence resides these days. This separation creates many interdependencies across the data center—and often, multiple data centers like EMC IT with facilities in the USA, Ireland, and Australia. The complexities that abstraction introduces, coupled with the dynamic and distributed nature of the environment, requires new technologies and new ways of doing things.

EMC IT like many data centers at one time relied primarily on manual hand-offs, approvals, and scripting for provisioning resources. The provisioning process took 10 days, with half the time just for approvals. To get to a private cloud model, it has had to automate and standardize processes and evolve a number of resource management functions in the organization.

EMC IT is not making these changes alone. Rather, it created an Enterprise Management and Services (EMAS) team and collaborates with business partners like VMware and others; much like any organization these days would look to technology partners, contractors, or other third-parties to augment core competencies.

The changes for EMC IT involve six (6) areas and include:

1.   Monitoring/Event Management  

Most compute, networking, and storage products these days do not include monitoring and management. When they do, these tools are specific to one product. To achieve a cross-domain view, EMC IT deployed EMC Smarts (formerly known as IT Operations Intelligence or ITOI). EMC Smarts is a deterministic availability solution with a powerful root-cause and business-impact analysis engine that monitors compute, networking, and storage to isolate issues. This comprehensive view greatly enhances IT effectiveness to identify and remedy problems to provide service assurance to EMC users.

2.   Automation via Orchestration

With a good understanding of overall data center health, EMC IT embarked on changes to automate provisioning VMware virtual machines (VM) using EMC IT Orchestrator to speed the back-office set-up work. The objective was to improve the internal customer experience.  This initiative is one of several steps to getting to infrastructure-as-a-service; by automating the VM provisioning process, EMC IT could present compute resources on-demand.

3.   Performance Management

While EMC products provide a single data-center wide availability monitoring capability, VMware vCenter Operations Management Suite was selected because of its ability to aggregate information generated by the many different EMC and third-party element managers for compute, networking, security, and storage. Future plans include leveraging additional VMware technology to drill down into the information provided by the different element managers to determine root cause on performance problems; much like Smarts provides root-cause determinations about availability issues. VMware vCenter Operations is a performance management complement to EMC Smarts service assurance for a complete picture of data center availability and performance.

4.   Capacity Management

Infrastructure capacity management is one of the longer-term initiatives. In the days when there was a direct relationship between applications, servers, and storage, over-provisioning fulfilled the need to meet current service levels with room to add capacity before it was really needed. However, this approach goes counter to the cloud imperative to lower costs. A private cloud model requires both on-demand resource availability and predictive analysis for effective planning for additional capacity. EMC IT is continually evaluating solutions for getting to a just-in-time model for data center resources.

5.   Federated CMDB

The dynamic nature of private cloud environments makes a configuration management database (CMDB) a necessity. This repository records the relationships between resources and applications spanning multiple data centers, and is a key element in getting to the services catalog required for all forms of IT-as-a-service offerings. EMC IT has implemented a CMDB, initially at the compute level and with plans to further populate the repository with information about networking, storage, and more.

6.   Governance Framework for Lifecycle of Tools

While many of the changes involved technologies, a key element in the move to private cloud has been the formation of a governance board to oversee software use. Applications and tools can be expensive. Similarly, the time and training needed to become proficient in new technologies adds expense to data center operations.

In its first year, this board was able to reduce the number of tools by 50%. Imagine the cost savings this provides in a FORTUNE 500 data center of the magnitude of EMC IT.

Well-Positioned for the Future

EMC IT is well-positioned to further abstract networking and storage, whether for creating services catalogs spanning a single or multiple data centers, or moving more intelligence to the control plane through software-defined networking or storage. It has made changes to monitor overall data center service availability and performance, speed provisioning, and centralize change management. An internal governance board has streamlined the software tools portfolio.

For more information about the EMC IT EMAS infrastructure management automation initiative, read the white paper: “Simplifying and Automating Management Across virtualized/Cloud-Based Infrastructures.”

About the Author: Mark Prahl