DISH CIO Atilla Tinic on innovating in the cloud

As EVP and CIO for American television provider DISH, Atilla Tinic leads IT strategy, delivery and operations for all of the company’s pay-TV and wireless brands—from the 25-year-old direct-broadcast satellite service DISH Network to the born-in-the-cloud IPTV service Sling TV and smart home service OnTech.

Now several years into the cloud journey, each of DISH’s businesses “is on a little bit of their own trajectory, when it comes to infrastructure and the cloud,” says Tinic. But the emphasis is on a cloud-first approach because, as he puts it, “all new technologies are being born in the cloud.”

CIO Contributing Editor Julia King sat down with Tinic at CIO’s recent Future of Cloud Summit to discuss innovation in the cloud, the importance of having a cloud mindset, staffing challenges, and lessons he learned along the way.

What follows are edited excerpts of that conversation. For more insights, watch the video of the event session embedded below.

On innovating in the cloud:

Making sure that we are keeping up [with the pace of technology change] is highly critical to me. But I do think the key is the cloud. And why I say that is all new technologies are being born in the cloud. And I would say all traditional commercial software vendors are trying to find their way to modernize their platform so they can be cloud native.

So it goes without saying that all our technology evaluations are happening there, which, by the way, is great, because we can evaluate capabilities quickly. We can typically do rapid assessments or spin up a sandbox environment to evaluate capabilities. So it puts us in a position where we can try it before we buy it and really learn about the platforms and kind of see what makes sense for us.

And I think this is critical because if you think about the rapid change of technology,… I do not want us to ever be locked into a single solution. I want to make sure we have an open ecosystem where we can replace components relatively easily as technology evolves.

On adopting a cloud mindset:

For network providers, there is a change in mentality of thinking . . . moving away from boxes and cables and that mindset, and really evolving into a world where we are adopting cloud software engineering best practices. You need to think about how you build your CI/CD pipelines, integrating with microservices and APIs in a config first approach.

We want to be in a world where we can purge and recreate any function on the fly, without high overhead of scheduled outages. So that is probably an easy example because from a carrier perspective cloud and your network in the cloud is cutting edge.

Even from a traditional IT standpoint, we still have paradigm shifts that we have to make. We want to avoid the trap of treating the cloud the same way we treat our data centers. In our data center, we will always over-provision to some extent, because we want to make sure we have adequate capacity and headroom so we can support lead times when it comes to procuring equipment. However, with the cloud, we can continuously optimize usage so that we can scale up as necessary. But also we need to make sure we continuously scale back down as we match the demand in our business.

On the cloud talent challenge:

Over the last couple years, we have worked closely with our partners to sort of co-innovate. And that has been a learning experience for all of us. There are a lot of capabilities that we have asked them to build into their solutions. We want to deploy changes faster. We want more visibility into performance. We want the ability to identify issues quickly. So, this creates a real demand on their part to put more cloud and platform engineering skills into their shops.

With consultants, we really lean on them to help bring thought leaders to us so that we can learn from them and also help us understand best practices. That allows us to evolve and improve. However, we also lean on them, to a certain degree, for elasticity of resources.

But most importantly, … is really our own employees. We are always on the hunt to grow our internal capability to support our cloud infrastructure. So, this includes cloud networking skills, cloud security, cloud architects, platform engineers. You name it, we want it…. I think the market has created some very big challenges. Because we are not the only ones who are looking for all these skills.

In this market, attracting and retaining [talent] is not enough. We have got to put the concerted effort into investing in our people, helping them retool with skills necessary to do their jobs, putting up people on paths to be certified cloud professionals.

7 lessons learned in hindsight:

Networking should not be underrated. If you are establishing direct connects to the cloud or  you have transit gateways, there is a lot of work that you need your networking team to architect and manage. Just make sure you are doing that well.

Build your environments with security in mind, and not as an afterthought. That is for the obvious reasons that you want to make sure that you have the right access permissions and firewalls, etc. to protect your environment. That is job 1. But it is also easier to move with pace when you have standardized security processes in place.

Automation. When you have automation, you start to force standards, basically, through the automation—every environment is not a snowflake. If you do that, that will not only help security, but it also actually helps your new development efforts move faster.

Look to standardize on tools. It is easy for every development team to pick something different, which creates challenges when you want to scale across teams. It is not always one size fits all, but make sure you have a catalog of tools you want people to use and it is not a free-for-all. You want to avoid tool sprawl.

Plan on how you are going to work in a multicloud environment. Even if you do not plan on being multicloud, you might find yourself there just because you are using SaaS components or you are integrating back to your own private cloud. So really think about your network, data, and security architecture as it pertains to multicloud

Cost. It is easy to have independent teams managing their environments, but I would really encourage some centralized oversight. When you are in a pure consumption-based model, it is easy for costs to get out of hand. So, you need to have that strong oversight.

It really just comes back to architecture and governance. It is like the old adage, “Measure twice, cut once.” If you spend the extra time up front, I think you will save a considerable amount of time and money in the long run.