Why telcos lag behind tech firms in the cloud, and how to choose a public vs. private cloud strategy
In a wide-ranging interview, Danielle Royston, one of the leading women in telecoms IT talks exclusively to VanillaPlus about the mistakes people make in their cloud strategy, why telcos are lagging behind tech firms in their cloud revenues, some bold claims, and saying Yes to more opportunities.
VanillaPlus: Danielle, you’ve been quoted as saying that network operators won’t fully benefit from the long-awaited “telco cloud” until they engage with the public cloud, including AWS, Google, and MS Azure. Why do you say that?
Danielle Royston: Moving to a cloud-ready solution on the ground in a private-cloud scenario doesn’t take advantage of one of the biggest reasons to go to public cloud — which is for the big cost savings that come primarily from the elasticity and auto-scaling of compute and database resources (pay for what you need, when you need them).
Customers who utilise a private-cloud strategy should not do it for the primary goal of cost savings. They should do it for the increased agility and increased ability to leverage their already purchased and provisioned compute resources.
I hear a lot of customers talk like public cloud and private cloud are the same — they are not. There is a big difference in terms of cost. In one scenario (private), you have to pre-purchase and provision all the compute and database resources, usually to 20-30% above peak capacity. In the other (public), you can provision for an average capacity and scale up or down as needed.
That’s significant in the dynamic world of telco.
VP: VanillaPlus recently reported exclusively on new research showing cloud revenues are growing much faster for tech firms than for operators. Why are operators lagging behind?
DR: Operators are lagging behind tech firms in cloud revenue growth because they don’t have the experience of deploying large-scale data centres like Google, AWS, and MS Azure. For example, our partner, Google, has spent US$31 billion (€27 billion) on building their data centres and investing in building their own private network. They spend more than US$3 billion (€2.64 billion) annually on security for their cloud.
It is hard for operators to keep up with the rate of innovation and spend compared to the big three. They have had over a decade to operationalise and derive the benefits of economies of scale, taking advantage of things like Docker, Kubernetes, plus new database technology like Spanner and Aurora.
Couple that with the price wars they are having against each other, and it really makes it hard for a customer to pick the operators over the big guys.
VP: Optiva is making bold claims about cutting telcos’ charging costs by 90% compared to Oracle, while boosting performance. Which of your customers have achieved that, and how?
DR: None are in production, but three to four are beginning POCs (proofs of concept) since seeing our TCO (total cost of ownership) business case comparison of what they are running today in private cloud vs. public cloud.
When we present our business cases — and we build them out in full, jointly with our customer’s input — even the strongest objectors can’t argue with our claims. It’s easy to prove. So, we welcome anyone out there that doubts us — let us come to you for free and we can prove it to you!
VP: Do alternatives to Oracle like cloud-based Google Spanner offer network operators other benefits besides licence cost savings? If so, what benefits?
DR: Every Tier 1 telco wants 10x the scalability of Oracle and every Tier 1 wants to reduce the US$10-100 million they pay Oracle. Lucky for us – Google Spanner delivers on both fronts. We recently completed a Google customer case study that talks about how differentiated Spanner is – Optiva: Transforming the telecommunications industry. But to be 10x more scalable AND 10x cheaper than Oracle, you need to be different. To solve this problem, you need to be:
- Distributed, but with strong consistency
- To solve this problem, you need thousands of servers concurrently writing — not 4, 10, or even 50 servers (like an Oracle cluster).
- Strong consistency is a hard problem to solve when there are a high volume of transactions — a lot of write traffic and a large number of distributed nodes (like there is in charging).
- Have global scale
- You need hundreds of thousands of servers, and they need to be redundantly located. The more locations the better if you want low-cost and still provide high availability.
- You can’t do this with 1 or 2 or 4 data centres — you need A LOT of data centres. This is because the more distributed the nodes, the more available the data and the lower the cost of keeping it highly available. Google’s scale enables the dramatic reduction of cost, while also making it a practically “infinite” scale.
- And remember, our customers are Tier-1 telecoms. They have great data centres. But even they don’t operate at Google scale.
- Have performance at scale
- This is where Google’s core technical differentiation really shines.
- Underpinning Spanner’s ability to solve this problem is the TrueTime API that enables distributed, strongly consistent transactions to operate at the maximum rate possible, given the capabilities of the network connecting the data centres.
- There are other databases out there, like CockroachDB, who are trying to replicate this idea without running in the Google Cloud. We think that without TrueTime and Google’s investment in data centres and networks, they just can’t get close to the performance and scale.
This innovation is what allows Spanner to address all of these constraints and deliver on 10X scale, while also being 10X less expensive.
VP: Are the telecoms and IT sectors closer now than 5 years ago to gender and race neutrality in their executive appointments? Or should Boards of Directors do more to encourage diversity?
DR: I’m not from telco. I came from high-tech which also has a diversity problem. But when I came into telco, I was surprised about how much worse it was than high-tech. For example, all of last year I think I met a total of two female executives in all of my visits to customers. Now, some organisations are doing better than others. Vodafone, for example, has a lot more senior women in key positions, so I believe there is progress being made.
As far as racial diversity, I don’t think there’s as much of a problem — this is a pretty global industry with a lot of movement in the senior ranks. So I see a lot of people from different countries moving to and within executive appointments. Within companies, the workforce tends to represent the people that live there, and I do see a lot of diversity within the ranks.
But where I see the most significant discrepancy is with women. They are missing. It’s pretty stark. I believe this reflects decades of women being left behind, not promoted, and then being discouraged and moving onto other roles or off-ramping their careers completely.
VP: I understand you were a Computer Science major at Stanford. What career or life advice would the Danielle of today offer to the Danielle who graduated?
DR: I would tell myself to try everything. Don’t say “no” to career opportunities that present themselves, even if the opportunities aren’t obvious as your next step forward. I’ve gotten a variety of experience not directly related to Computer Science, and it’s helped me to become a better leader and CEO.
For example, for years I thought the eight years I spent in HR, when my goal was to become an SVP of HR, was thrown away when I was asked to become a turnaround CEO of a software company in 2009. But I was completely wrong.
Since software is a people business where your most significant expense is compensation, learning and knowing how to manage human capital may be one of my best experiences. Today that serves me more often than my Computer Science degree from Stanford!
If I had said no to that initial offer to move into HR, I doubt I’d be a CEO today. Same goes for my time in sales or the time I spent managing IT.
So, say YES more to opportunities — it almost always works out for the best.
Danielle Roston, Optiva’s CEO, was talking to Jeremy Cowan (left), editorial director of VanillaPlus and IoT Now.