Broadcom’s Andy Nallappan on what cloud success really looks like

Companies moving to the cloud often find themselves at a crossroads near the midpoint of their migrations, spending more than they intended and getting out less than they hoped.

Often that’s because their IT organization isn’t equipped with the culture, mindset, and skills necessary to capitalize on the cloud.

Andy Nallappan has had a long career in IT, including CIO roles, but his current job at Broadcom is managing the company’s external cloud platform, DevOps, and SaaS operations across multiple software divisions. Along the way he has developed deep expertise in what it takes to succeed in the cloud.

In this interview, Nallappan has much to share on dynamic cloud architecture, cloud cost containment, developing a cloud-conscious culture, and why it makes sense to give engineers financial goals when giving them ownership over their cloud estates.

Martha Heller: Since joining Broadcom 2016, you have been CIO, then CTO and head of software business operations, and now CTO, chief security officer, and head of software engineering and operations. How has your role shifted through that trajectory?

Andy Nallappan: I went to school for mechanical engineering in the 1980s, when people were just starting to use computers. As computers evolved, so did I. I got into software engineering and spent the next 30 years in IT, including roles as CIO of Avago Technologies and Broadcom, which involved a lot of acquisitions integration.

A few years ago, I agreed with Marc Andreessen that ‘software is eating the world’ in industries from banking to government, retail, and healthcare. These businesses had traditionally transacted with end customers through four walls. But over the last three to five years, that has changed. Now customers dictate how, when, and where they do business. The paradigm has shifted: The four walls and time zones have come down and customers can transact 24×7.  

With this change, technology has become a major part of revenue beyond CRM and sales enablement; technology has moved into the line of business [LOB].

LOB leaders want to control their own destiny and not be at the mercy of IT. They are saying, ‘Give me the technology budget and I can decide where to innovate,’ and CEOs and CFOs are saying, ‘If LOB leaders will commit to profit, well okay, let’s let them have skin in the game.’

I saw this shift happening at Broadcom, so I decided to shift out of IT and into software operations.

What is your focus as head of software business operations?

In our hardware business, the COO runs the supply chain. My role in the software business is similar. I manage the external cloud platform as a partner to the businesses. My job is to liberate the engineers in the businesses to do incremental innovations and not boil the ocean.

At Broadcom, each software division is run by a GM who is responsible for the P&L. I work across all the software divisions and manage three things: 1. DevOps, the tools used to build the software; 2. the cloud platform, which developers build the software on; and 3. SaaS operations, which runs that. Since the divisions are all using the same platform and tools, we can leverage common software operations so that customers have a similar experience. We have multiple divisions but a common toolset and platform.

Software is software, and while the user experience and transactions layer might be different in each division, there are commonalities in the layer below. As head of software business operations, I manage the noncore elements of the software so that the LOBs can be free to manage the core. My role is part of the revenue, not the cost.

You also manage Cloud FinOps to tackle the problem of overspending, which is a problem that many companies face when they first move to the cloud. Can you explain?

When the cloud is operationally new for a company, they spend too much money without achieving the right outcomes. The biggest mistake you can make when you move to the cloud is not changing the culture. Operations in on-prem is completely different than in the cloud. If you run your operations the same way as when you operated in data centers, the cloud will be three to five times as expensive. There is no question about it.

With on-prem, the mentality is around allocation and sunk costs. In this model, software engineers are not motivated to keep costs down because they think of the allocation as a tax. ‘My LOB paid the tax, so it doesn’t matter if I waste money.’

But in the cloud, that mindset needs to change. Software engineers should know how much the cloud costs for their area, have visibility into where they spend money, and be equipped to manage those costs, because the cloud is a land of opportunity. Allocation doesn’t work in the cloud.

Can you walk through an example of FinOps in action?

Our on-prem software testing labs were running different operating systems and hardware combinations, because every customer is unique. We had to test each combination, and we were allocating those costs to the software businesses.

When I moved the labs to the cloud, I told the engineers to move only what they are working on right now. When they need more cloud services, they are empowered to get them; they don’t need to ask my permission.

But I also gave them financial goals and visibility into the costs of the tools. I do control the costs. I govern the costs. When we were in data centers, IT would buy the hardware, and because they knew they needed headroom, they would overbuy and allocate those costs to the businesses. Today, the engineers themselves are empowered to buy only what they need, and to hit their own financial goals. Those savings are a part of the P&L and have a positive impact on the engineers’ bonuses.

FinOps has helped to change the sunk cost mentality. The engineers know that if they bring cloud costs down, they are contributing to their business’s P&L and their bonus goes up. They are excited by the new model. ‘As long as I save money and align with my business model, I can use the cool services in the cloud. I don’t have to ask permission.’ 

What are other mistakes companies should avoid when moving to the cloud?

Companies should avoid the lift-and-shift approach and understand that you cannot refactor every application on day one. To rightsize your move to the cloud, you need people who know the cloud, but these are not the same people who have been managing your on-prem environment for the last 20 years.

It is also important to understand that in your first year in the cloud, your costs will be higher than in years two, three, and four, but you will be able to do incremental innovations. With on-prem data centers, you innovate once in five years because that’s how data center technology refresh works. With the cloud, you innovate every quarter, month, and week. That’s the beauty of the cloud.

What is your advice to companies moving to the cloud?

Moving to the cloud is all about changing the culture, rightsizing your approach, and having a roadmap to get into a dynamic architecture that scales up and down during the week, and comes way down on the weekends when people are not working. That’s how you can scale.  

But you cannot put the same wine in the new bottle and call it new wine. You need to develop a cloud-conscious culture by bringing in new leaders who have a FinOps mindset and a focus on security. It is also important to bring the software groups together. We have consortia for DevOps, security, and cloud, so that the developers can talk to each other and not reinvent every time. Developers love to build their own tools, but that’s not necessary in the cloud.

We bring the developers together so that they can be proud of what they are doing and share it. With FinOps, we shift their pride to saving money and contributing to customer outcomes. It’s not about showing them a PowerPoint that says, ‘Cloud Conscious Culture.’ It is about showing them the outcome of their work.