X

VMware on the cloud and virtualization (Q&A)

VMware cloud executive dishes on how the company thinks about private cloud, virtualization, and what problems customers need to solve.

Dave Rosenberg Co-founder, MuleSource
Dave Rosenberg has more than 15 years of technology and marketing experience that spans from Bell Labs to startup IPOs to open-source and cloud software companies. He is CEO and founder of Nodeable, co-founder of MuleSoft, and managing director for Hardy Way. He is an adviser to DataStax, IT Database, and Puppet Labs.
Dave Rosenberg
5 min read
Chris Knowles, VMware Enterprise Cloud Architect
Chris Knowles, VMware Enterprise Cloud Architect

VMware has come to define the virtualization market with more than 190,000 customers and dominant market share. But does that success translate to the cloud world?

To better understand how VMware thinks about virtualization and cloud, I talked with Chris Knowles, VMware's enterprise cloud architect, about how the company approaches cloud and what IT feedback the company is getting from customers.

Q: While some organizations have gone all-in on adopting cloud technologies, many organizations are still planning the best approach. How are the companies you are working with looking at implementing public, private, hybrid cloud environments?
Knowles: The approach that we discuss with enterprises is to essentially take the same path to the cloud that they did with virtualization: start with the low hanging fruit that can be well served by the cloud model and show clear, rapid results.

Customers can develop, test and then move tier 2 and tier 3 applications to a private cloud. These are great ways to get started within the cloud. This allows you to learn how to best leverage the technology as it relates to your specific needs, while minimizing the risk exposure to the enterprise. As IT matures in its understanding and use of cloud-based compute services, they can expand to more strategic use cases.

Starting with a private cloud model helps many enterprises achieve the benefits of cloud computing without tackling issues like security that many people associate with public cloud solutions. For some enterprises, challenges around compliance and risk may keep them from ever moving to a public cloud strategy. For others, gaining a solid foundation and understanding of cloud through a private cloud model inside the walls of their datacenter will help them augment their overall resource model with public cloud resources.

VMware obviously has a very strong market position in virtualization, but virtualization is not cloud. How does the company think about addressing that fact?
Knowles: That is absolutely correct, virtualization is not cloud. However, without virtualization there is no cloud.

To VMware and the broader industry, virtualization is the necessary foundation upon which the cloud is built. With virtualization we set out to replicate the physical characteristics of a server in software, CPU, memory, storage, network devices, etc. With cloud computing we take the flexibility that is offered by a virtualized environment, and on top of that abstract away the constructs that came from the physical world and layer in additional capabilities that are not inherent in a virtualized infrastructure.

Without virtualization that is robust and feature-rich with capabilities, such as on the fly automated workload rebalancing and policy-based networking in software, the promise of cloud's agility and elasticity is reduced to marketing hype. In the cloud, users no longer concern themselves with things like where their workloads (VMs) are running, or what storage they are consuming. They simply are selecting where their workload should run from a class of service, and the cloud takes care of the rest.

In a virtualization-only environment you still have the virtualization administrator creating virtual machines and providing those to the user. In contrast, in the cloud we have a self service model where the end-user can select their virtualized services or applications from a service catalogue, what class of service they need and deploy the resources on their own.

As the organizations with which you are working implement cloud infrastructure, what are some of the management challenges they are seeing?
Knowles: The challenges faced by organizations when adopting cloud really fall into two categories: process challenges and technical challenges.

Cloud computing really turns the process of delivering and managing IT services relative to current practices upside down. With traditionally delivered services, whether physical or virtual, they tend to have very rigorous processes for acquiring machines, moving those machines into production and managing changes to the infrastructure once it is in place.

Mature enterprises tend to have processes akin to ITIL that are very structured and have a lot of steps to get from the end user thinking to themselves "I need a server" to actually being given a user ID and IP address of a system they can go and login to. It's not uncommon for those steps with all the associated routine approvals to take weeks, or in many cases, even months.

With cloud, the delivery chain, from thought to login, can be delivered in a few minutes. The speed with which an end-user can satisfy their requirement for a new system in the cloud is one of the single most compelling reasons people want to move to the cloud. This is not simply convenience, but a real enabler of faster time-to-market and greater freedom to innovate.

How does managing cloud and virtualization differ from managing physical resources? To what extent should the approaches be integrated?
Knowles: At the end of the day, whether you are running your VMs on a virtualized infrastructure or in the cloud, they're still running on top of physical infrastructure. Enterprises need to manage each element of their resources to ensure services with the performance and availability that the organization demands.

Tools must understand and embrace the dynamic nature of virtualization and cloud, and their successive abstractions from physical infrastructure. Faster, more granular element inspection isn't the goal, but an API-based approach that focuses on the end-user service being provided in user-centric terms (an SLA, for instance).

Third-party vendors such as Zenoss have announced unified management solutions for VMware. Does the ecosystem matter, or does VMware ultimately want to own every part of the infrastructure?
Knowles: The vendor ecosystem is essential to the success of cloud computing and the success of VMware's ability to execute on our vision for cloud computing.

Cloud computing is a fundamental paradigm shift in computing and will have huge implications for the future of IT. The challenges created by this paradigm shift are too broad and too deep for any single vendor to address. A very rich and active vendor ecosystem focused on cloud has already emerged, and will only continue to mature as cloud computing matures in the industry.