The challenges of scaling the edge, with Mark Howell, Ford Motor Company

The challenges of scaling the edge, with Mark Howell, IT Facilities Planning and Engineering lead (EMEA), Ford Motor Company

Often, we struggle to discuss the edge – one of the IT world’s hottest trends – because it’s difficult to objectively define it. “The edge means something different to every person,” says Mark Howell, of the Ford Motor Company. Howell oversees the construction of every new IT facility the veteran automaker builds and is the lead for EMEA region design, planning and engineering.

From his perspective, the edge is effectively distributed technology, and need not encompass servers, storage and switches. By that token, Ford’s first edge site wasn’t a micro data centre, but the first remote offices and factories that Henry Ford built all those years ago. The company’s Paris office opened in 1908, the Kansas City assembly plant opened in 1911. By the end of the 1920’s Ford had more than 20 overseas assembly plants.

“The distributed technology has changed but the principles remain, to improve design, manufacturing, distribution and logistics, sales, financing, servicing and customer relationships. And to reduce cost so that the cost of the product can be reduced for customer benefit.”

Back to basics

Mark’s definition keeps things simple. If a site is connected to Ford’s “core” (enterprise data centres located in North America), it is an edge site. Ford operates more than sixty factories around the world plus the joint ventures it has with other carmakers. Factories are not the only sites upon which its global operation depends. There are product development centres administration buildings, offices. All of these are connected to Ford’s network. All of these “are edge sites.”

These days, of course, the technology that powers Ford’s edge is not just machines and energy, but IT: routers and switches for the wide area network and local area networks (not to mention conferencing equipment, digital displays, desk phones, CCTV and so on). For servers and storage, a site might be connected to a regional or local data centre for latency or data residency requirements — each needing cooling, power (this can include a UPS and a backup generator). Each site has its own requirements that must be understood.

“Keep people away”

One of the biggest issues the edge throws up is volume and scope. Designing, constructing, operating and maintaining a sprawling portfolio of locations is a challenge that gets more complex with each additional site. The biggest challenge? “People!” — more sites mean more people for facility and design engineer leads like Mark to coordinate, accommodate, and, sometimes, confront.

At the design stage, it is common for there to be disagreements, and there can be varying levels of “knowledge, skills, experience and competence”. Every facility also needs the approval of “gatekeepers”, such as local government or municipal officials. There’s no guarantee planning or building control officers will understand best practices or be well-versed in the particularities of construction, such as the differences between traditional bricks and mortar, post and beam, modular options. Aesthetics can also be an issue — shipping container based solutions are not something local municipal Planning officers are keen to see.

When it comes to construction, operations and maintenance, the same hurdles crop up. Design and engineering leads have to contend with mistakes that cause partial or complete outages. It is the “the focus on operating cost” which is often the cause of downtime, says Mark. For maintenance, Mark adds that he is often at the “mercy of white van man”:

“Maintenance must be completed as described in the original equipment manufacturer service manual, this is the same principal as Ford require customers to do,” he explains. “The maintenance work must be completed by people that are trained and certified. The guy that says… “It just a UPS, I can maintain that for you and much cheaper than Schneider Electric” is the guy to avoid. All UPS fires I’ve investigated have had a root cause of poor installation, poor maintenance or no maintenance”.

His advice for the growing number of companies scaling their distributed IT portfolio is to “understand the location, understand the need, and keep people away!”. Establish trusted technicians and be strict with those who you grant access. For instance, steer clear of suppliers who scatter when the going gets tough. “Some suppliers don’t like to take on work that they see as risky. For example, civil works. However, they are happy to sell you a modular DC though!”

Don’t miss Mark’s session at Data Centre World London, where the data centre veteran will share lessons from his decades of experience building at the edge. March 11-12, ExCeL London. Registration Free.