Internet Boss. ‘Planet earth is now a data centre’



Jay Adelson

The internet – created in the 80s, but the world wide web officially went live in 1991 and has since then grown into a beast with around 3.2 billion of the worlds population currently online.

Internet guru and entrepreneur Jay Adelson has witnessed the explosion of interconnection labelling it as the hub and central point of the internet.

His view is that we need to have a more decentralised model of the internet – and fast, to stop this force from getting out of control! Revealing a unique take on how companies should be making gender diversity a top priority – Abigail Opiah has more.

In 2014 Adelson co-founded Center Electric with Andy Smith.

In 2013 he founded Opsmatic, a technology company that improves productivity on operations teams, but before all of that, Adelson was the mastermind along with co-founder Al Avery who both joined forces in 1998 to create global colocation data centre company Equinix.

What was your journey like from studying film and broadcasting to becoming an internet entrepreneur?

 When I was in school and I was studying film, I had an obsession with communication and this view that there was this incredible power that film had to bridge cultural gaps and connect people together.

I was also an internet nerd; I was online when I was 11 years old back in 1981.

Vincent English, CEO of Megaport, a company where Adelson sits on the board

While I was in film school, right around the time that the Tiananmen Square protests happened, I remember reading about the fact that email in its earliest form, was the trick that students used to organise for that protest.

It made me realise how it can cross-gaps, and what the power of a truly global internet could create.

When I moved to California to work in sound and post production, I was still very much fascinated with spending my time online.

A friend was working for the very first internet provider in the U.S., and he asked me to join, and at that point, nobody cared about the internet so there was no specific training that was needed other than knowing about telecommunications and understanding the basics of how it worked, so I got lucky.

I went to work for this company and was able to spoon-feed myself slowly with all of the knowledge I needed to be a network engineer and grow up in the 1993 days of the internet.

What led you into the data centre industry?

 I was running network operations at Netcom, and in 1994 when the internet went commercial and stopped being something that was only operated by research institutions and the government, there was an effort to commercialise it and hand it off to private industries.

Netcom was already online, and we were struggling to scale our interconnection with other networks and services, because there was a hierarchy where you had to pay your competitors to exchange traffic with them.

This was frustrating because at the time we were doing pretty well, so to have problems where your customers see your services as not performing well because your competitor is essentially a continued gatekeeper to your ability to scale, was very uncomfortable.

I got a call from the network systems laboratory at Digital Equipment, and they said to me that they had a concept that they wanted me to build out, where they wanted to create neutral interconnection between network services and enterprises.

I thought that sounded amazing so I quit my job and went to go and work for Digital Equipment and we built the Palo Alto Internet eXchange.

When we built that exchange, it was unavoidable that it had a data centre component to it, because we had to have a place that was secure, air-conditioned and designed for all of the high-density network gear that was going to come in.

We had to build a data centre around an interconnection point, and we learned very quickly that everybody wanted to be in the interconnection point.

It was extremely successful and it is still operating today as one of the largest exchange points in the U.S., and it was doing so well with the data centre interconnection combination that we spun it out and created Equinix.

That is how we ended up becoming a data centre company – it came from interconnection.

What made you join Megaport?

 If you think of interconnection as a hub of the internet – the way that these facilities and these interconnection points operate, data has to flow to the most efficient path it can find.

If you look at the Equinix model as an example, the cross connect inside the data centre, was the way that an enterprise had to connect with all the services.

However, as hyperscale started exploding, and became more prevalent and more distributed, we started getting things like data privacy and local verses of remote data, and hybrid clouds.

Planet earth is now a data centre needing an interconnection fabric like cross connect for the earth.

It’s no longer in the four walls of a data centre, thus Megaport became interesting to me because I feel as though it has created that interconnection fabric for the planet earth.

The promise of this software-defined network that we were all hoping for 10 years ago, is finally coming to its fruition with Megaport.

Also, speaking as someone who has started API companies like Opsmatic, people understand API’s better now, even small enterprises understand the use of micro services and now everyone is into SaaS and PaaS, so the revolution around API’s allows people to connect in ways that they could not do before.

The other issue that forced people into these big data centres was capacity and bandwidth.

This is now not an issue – nobody talks about there being a limit of fibre capacity or wavelengths.

Lastly, it’s really down to Amazon, Microsoft and Google as well as all these other services that have launched their direct connect services to support the enterprise needs and then put them all over the earth – that changed the game too.

What are the two most exciting projects you’re involved with now?

 One of the things that fascinates me about Megaport is their portal platform – I wanted to get involved being part of the innovation committee that is internal to how Megaport operates.

What they have asked me to do is look at their existing infrastructure and portal platform; to get involved in the feature set and what kinds of changes will be appropriate as we look at new distribute applications and new customers as Megaport expands from the fortune 500 into the unfortunate five million.

There is no question that it’s incredibly easy to provision, and it’s incredibly easy to use their product, but architecture has become more complex.

Data privacy and its laws are changing, thus Megaport has to role with it, so I want to be involved with that and help them through it.

Secondly, Megaport as a company is moving into new markets.

A typical software defined service may have around five or six interconnection on-ramps just in the U.S., whereas Megaport has 115 of these on-ramps around the world.

That’s the other piece that’s really interesting to me because back when I was in my Equinix days, having 100 points or on-ramps around the world is something I couldn’t even imagine.

They have gone from no network infrastructure, to having major network infrastructure in just five years.

With the World Wide Web turning 30 this month, how have you seen the evolution of internet over the past 30 years, and what could be improved?

 While I am incredibly optimistic with where the internet is going, and I do believe that it is an incredibly powerful force for good, I do tend to side with Tim Berners-Lee and his view of where it needs to go.

There is no question in my mind that a few large companies have centralised a lot of the use of the internet, and as a result, there has been some negative outcomes.

Not just around data privacy but if you look at things like the elections in the U.S. in 2016, it’s very clear to me that we need to have a more decentralised model of the internet.

To do that is going to require a number of changes both on the technology and governance side.

I am watching Tim’s efforts really closely, and I am fascinated with where we could go as a self-governing model.

I have three children who use the internet as part of their everyday lives, and I have seen the great things that have come from it like Wikipedia and the constant ability to connect.

I have also seen the negative side, so I really do believe we need to stay in front of this and put some new models out there to fix these problems.

What do you expect or predict from the next 30 years?

 I think we are going to stop calling it the internet.

In 30 years from now, possibly even three years from now, we are going to seize to think of the internet as a platform, but more so like the air we breathe.

Things like provisioning, establishing connectivity, or even the interface, is all going to be integrated into everything that we do.

If you look at things like voice recognition, AI and autonomous vehicles, it’s all being integrated and the internet just becomes like electricity.

We have almost stopped thinking of it as a thing.

The internet is about best effort, so for an enterprise, they need to have something that is outside the internet that is a guaranteed connection, thus in three years and certainly in 30 years, you will have this combination of public and private interconnection.

As the co-founder of Equinix, how have you helped to shape the company over the years?

 When we created Equinix, myself and my co-founder, Al Avery who passed away in 2009 – both had a vision around interconnection.

While other companies were mistakenly chasing things like valuation, hosting, and basic bandwidth, Equinix stuck to the vision.

We stayed focused on interconnection, and when there was Nasdaq crashes, financial recessions, and telecoms collapsed in 2001, the one company that survived all of these changes in the industry was Equinix because we focused on interconnection and not just on real estate.

That was my role in the company, both at the beginning and throughout the years as an adviser.

I always held the torch that interconnection was so critical, and I think they have done a very good job until this day.

I don’t work for the company now, but I have still maintained great relationships with the team over there, I have certainly been applauding their success.

Would you start another colocation company?

 Never say never! Today the value of interconnection has expanded beyond the four walls of a colocation facility.

There is a great balance and symbiosis that occurs between interconnection and colocation, where the interconnection services add value to the colocation services and you can’t have interconnection services without the colocation.

Arguably, there are enough colocation companies, and I think I am more interested today in the services than in the four walls itself, but I still love colocation and enjoy advising those types of companies as much as I can.

How do you see the colocation market transforming?

 There is a shift in colocation, and the focus becomes less on the single giant centralised facility, and moves more to the edge and local facilities.

The need for the applications have also changed – if you look at blockchain and how it works, the nature of applications are more edge savvy and are architected that way.

This means that people have a virtual presence in more locations.

There is this big trend around virtualisation that has been talked about for years where people have moved towards virtual continued servers, yet the network has never been virtualised.

As the servers became virtualised, the idea of quickly deploying your virtual presence without physically visiting a colocation centre meant that now it made more sense to be able to hit a few buttons and launch a bunch of instances in hundreds of locations at the same time.

Now the same thing is true at network, because now networks have become virtualised in a way that the servers have with Megaport.

When this happens, colocation changes, because then colocation becomes a different product where you no longer have to visit it and the density increases because you get to use commoditised equipment, and architectures that are more shared.

It’s almost like a pendulum has shifted – originally everyone was outsourcing to large companies like IBM to manage their services, then the pendulum swung back and everyone is buying small cages and colocation facilities.

Now the pendulum has swung back the other way again, where we are using all these virtualisation services like Megaport.

The good news is that they have more revenues per square foot, and you can build them in places where it would have been impossible to do so in the past because you can build them in smaller locations, or even vertically.

What are your views on markets like APAC, Africa and LATAM?

These markets are where investments need to be made – look at the Africa market for example, back to my point about trends going more local, and the strengths of every market now offering their service as a basic on-ramp, you are going to see more investments in the Africa market, and more colocation in those locations. Microsoft just launched in South Africa, which was major.

If you’re growing a market, you need to be able to support that with services like Azure or AWS, and you can’t backhaul it to Hong Kong or even the U.S., you need that data centre capacity in order to support that growth.

In South America, there will be a tone of investments in places like Brazil. There is a huge investment opportunity in the LATAM market, and when I talk to data centre operators like Digital Realty and Equinix, that’s where I see them making future investments.

You see that they change their products to be able to support hyperscalers as a product in these markets. It is the same story for the APAC market – data centre operators are expanding as fast as they can there. Singapore is a fantastic example – with huge investments being made by cloud providers.

Japan was doing a lot of backhaul for a long time, but now has really embraced how the cloud operates. There is huge move towards that local infrastructure.

With the recent outages at Google, Facebook, Instagram and WhatsApp, do you think the internet is resilient enough for all the promises of a better connected world?

I do believe that there is a path towards a better connected and secure world, but I don’t believe that today it is showing the resilience it needs. We are in a dark period, where these central control points and large powerful entities unintentionally created these security problems.

The most important first step we have achieved is self-awareness. The question is how long is it going to take for us to reinvent ourselves on the platform. The internet is an open road, and what we put on it is completely up to us. So is the internet resilient enough – yes! It has the reliability, capacity, capability and the redundancy to accept new protocols.

Have we deployed what we need in order to secure it, the answer is no. The internet is so open and there is so much that we could be doing with it, but it is very hard when it’s controlled by these profit interests.

The challenge is also getting to a stage where everyone feels confident and regains trust, which I think will take years. I was the CEO of a company called Dig, which was a large consumer internet site in the mid-2000s, and we drove all sorts of traffic, but part of the issue that we had was identity management and guarding against things like fake news. We were thinking about this in 2005, and we didn’t fix it by 2010 when I left the company.

What are the biggest socioeconomic risks that cybersecurity poses to us?

It is a multi-front war. Even small or medium enterprises have to have people thinking about these issues every day. There is going to be a period of time where it will still be inaccessible from a technology standpoint, and it’s still too hard for that size of business to have that knowledge in-house.

Therefore, what is happening is that the ability to deploy that security across hybrid environments is becoming easier for these smaller businesses and consumers to take advantage. That is going to take years unfortunately.

Cybersecurity experts you talk to will tell you that they have the architecture right now, but small and medium companies cannot even afford to talk to these experts on the phone. There is also the issue of cost creating a gap in terms of the deployment of some of these technologies.

With all the knowledge you have of the industry now, what message do you have to your younger self?

If I could go back to the 90s when we were building some of these infrastructures, we should have pushed a little harder when there was less interest that was fighting against us, because nobody really understood where the internet was heading to back then.

Some telecom companies back then didn’t even sell internet services, whereas now, everybody does. We could have pushed harder back then for better net neutrality, and gotten it locked into place before things got out of hand.

We were so focused on the commercial problem and how to scale the internet, we weren’t really thinking about policy as much.

I wish I could go back to my 27-year-old Jay and say get more involved in policy, because now I feel like it is unfortunately almost impossible to correct.

Do you think the efforts made in the tech industry towards gender equality are enough?

One of the reasons why I got into ventrure capital in the first place because I believed that there was not only a selection bias, but also a huge corruption around power bias that controlled indefinite in Silicon Valley.

There’s exceptions to that rule, but when you have the same group of old white guys making all the decisions, I don’t think it was shocking that there was all this bias behaviour. #metoo certainly exposed many of the challenges, but those of us who have been working with women in the industry for a long time, felt that there was an inequity and beyond that, abuse.

This is an incredibly disappointing and sad state that we are in. What we need to do going forward is follow the money.

If you trace investment dollars all the way back to where it comes from, what I’ve seen, which is a positive change, is it creates qualifications around investments that say ‘we will not invest in a firm that does not have a diversity story which is compelling and real’, and not just quota.

That is happening right now, it is slow but it’s happening. Then what happens is, the VC’s bring in women into their general partnership. As a VC, I saw thousands of pictures of female entrepreneurs coming through Silicon Valley, so I have definitely seen an increase in the interest in female entrepreneurs compared to ten years ago.

Part of the reason why that is, is because there is this sense that there is a less discriminatory process by which they can pitch and get investment. There is another set of questions around these companies and how they managed diversity in the workplace.

We’ve seen some terrible examples in Silicon Valley of how that has been managed, but the good news is that the spotlights are on it.

Large powerful companies like Facebook, Microsoft and others have implemented new policies that improve diversity management in the workplace.

The old HR methods are falling short. I have seen some interesting new applications that HR departments are embracing to create feedback loops in real-time around treatment, sexism and harassment. However, we have a long way to go!