Zum Hauptinhalt wechseln

 Subscribe

Supporting a network in transition: Q&A blog series with David Lef

In a series of blog posts, this being the third, David Lef, Principal Network Architect at Microsoft IT, chats with us about supporting a network as it transitions from a traditional infrastructure to a fully wireless cloud computing platform. Microsoft IT is responsible for supporting 900 locations and 220,000 users around the world. David is helping to define the evolution of the network topology to a cloud-based model in Azure that supports changing customer demands and modern application designs.

David Lef explains the planning and processes behind migrating to a wireless networking environment, including primary drivers, planning considerations, and challenges.

Q: Can you explain your role and the environment you support?

A: My role at Microsoft is Principal Network Architect with Microsoft IT. My team supports almost 900 sites around the world and the networking components that connect those sites, which are used by a combination of over 220,000 Microsoft employees and vendors that work on our behalf. Our network supports over 2,500 individual applications and business processes. We are responsible for providing wired, wireless, and remote network access for the organization, implementing network security across our network, including our network edges. We make sure that the nuts and bolts of network functionality work as they should: IP addressing, name resolution, traffic management, switching, routing, and so on.

Figure 1. The Microsoft IT environment

Q: What is driving the shift toward a wireless environment?

A: For Microsoft, it’s two main things: first, our employees want flexibility in the way they get their work done. Our users don’t simply use a workstation at a desk to do their jobs anymore. They’re using their phone, their tablet, their laptop, and their desktop computer, if they have one. It’s evolved into a devices ecosystem rather than a single productivity device, and most of those devices support wireless. In fact, most of them support only wireless. The second motivator is simple cost effectiveness. It’s cheaper and simpler to set up and install a wireless environment than it is to do the same with wired infrastructure. It also makes upgrades and additions to the networking environment easier and cheaper. With wireless, there are no switch stacks to add and no cables to run.

Q: How did you begin planning for this?

A: When Microsoft started accepting and supporting mobile devices connecting to the corporate network, it was clear that the way our network was accessed was going to change. We initially planned to provide wireless support to the physical locations that needed it the most as a support for our wired infrastructure. However, traffic and use analysis showed that the wireless network was very quickly becoming our main network infrastructure, from a user’s perspective. We knew that wireless needed to be there to support mobile devices, and we knew we had to plan for the wireless network to support most of our end-user connectivity, eventually. We looked at the device profiles across our information worker roles to assess what was necessary, and we built out a network to meet that demand and make sure that it scales well with future growth.

We had, and still have, a lot of wired infrastructure that simply isn’t being used to its potential. At many of our information worker sites, wired port utilization is less than 10 percent. If you average it out across all of our user sites, it’s closer to thirty percent, but when you do the math, it still ends up being a lot of investment in network infrastructure that simply isn’t necessary. Over seventy-five percent of our sites are targeted for wireless-first, and we’ve been going through the process of removing dependencies on the wired network infrastructure from a user perspective. In some cases, that means putting wireless network adapters into desktop computers that don’t natively support wireless, and simply making sure wireless connectivity is enabled and configured on those devices that do support it. The more complete we can make the transition to wireless in terms of number of devices, the sooner we can retire the existing wired infrastructure and realize the cost savings from it. We estimate that our wireless-first strategy will result in a reduction in network equipment of more than fifty percent.

Q: What are your key considerations in this project?

A: It’s driven primarily from the high-level goal of cloud first, mobile first. Wireless networking simply complements both of these strategies; it’s a logical and necessary part of the larger puzzle. We are a business, of course, so cost and capital cost savings are important. Migration to wireless as our primary network infrastructure means long-term cost avoidance, less equipment to buy, and decreased maintenance requirements.

We also want the transition to be as non-intrusive as possible to our users. We’re going on-site to make sure they’re ready for the transition to wireless. This might mean helping users install or configure wireless adapters and showing them how to perform tasks, such as installing an operating system, differently. We also want to educate them about using the network and get them comfortable with being their own first level of support and solving basic issues they might encounter.

Q: What have been or will be the biggest challenges in making this work?

A: We’ve run into some challenges in a few different areas. Different devices and their drivers have their peculiarities and issues, whether that’s with a new wireless adapter we’re putting into an existing computer or access and authentication mechanisms for devices that use older wireless network hardware. We also have a lot of wireless access points around the globe, so standardization of those access points has been a challenge. With the advent of bring your own device (BYOD) and the emergence of the “Internet of Things” (IoT), many more wirelessly networked devices are showing up in our environment, and bandwidth is always a concern. A big part of managing this trend is realizing that not all IoT devices need to be included in our corporate network—only those that will benefit from the functionality that the corporate network enables. We’re providing the highest level of wireless bandwidth that we can, as far as supporting devices and meeting transmission standards, but we’re still closely monitoring bandwidth availability to ensure that we’re eliminating any unnecessary bottlenecks.

We’ve also had to address some changes in processes and conceptions. In some cases, older technology that’s in use doesn’t work with wireless, so we have to show users how to do tasks differently, or give them an alternative method.

Q: Is the technology available today to make this successful?

A: Yes, and we’re in the process of rolling out 802.11ac, which gives us more capabilities and bandwidth across our wireless infrastructure. We’ve also committed to having 802.11ac fully implemented before we begin any mandated removal of our existing wired infrastructure. We want to ensure that our wireless network can provide our users a satisfactory level of reliability and performance before we start removing the old way of connecting.

We’re continually rolling out upgrades and changes to our infrastructure to implement 802.11ac, but it also means making sure that existing equipment that our users employ isn’t being removed from the network inadvertently. Whether we provide an 802.11ac-compatible solution or simply replace the device itself, we’re very conscious of reducing the negative impact of the change on our users.

Q: What is the roadmap for pilot and implementation of this project?

A: It’s already in place and underway. The pilot project has been closed and we have 660 sites targeted for wireless infrastructure updates and conversion in the next 24 months. The other 200 or so will retain wired functionality—these are datacenters, engineering centers, or locations where our customers or users might still require wired connectivity. In the grand scheme of things, we’ll be cutting over 90 percent of our end-user network infrastructure. Wired ports will still be available where they are needed, but our footprint and associated resources needed to support it will be massively reduced.

Learn more

Other blog posts in this series:

Learn how Microsoft IT is evolving its network architecture.

  • Explore

     

    Let us know what you think of Azure and what you would like to see in the future.

     

    Provide feedback

  • Build your cloud computing and Azure skills with free courses by Microsoft Learn.

     

    Explore Azure learning


Join the conversation