When Hong Kong-based telecoms company PCCW Global delved into the world of Silicon Valley VC-based tech start-ups to acquire a company called Console Connect in 2014, it is unlikely that too many cloud service providers or users took too much notice. But now the term multi-cloud has established itself, many of them may well be glad that it happened.
PCCW Global owns and manages what it claims is one of the most diverse global fiber networks in the world, enabling it to operate a global IP and MPLS network that connects cities across Asia, Australasia, Middle East, Africa, Europe and the Americas. The acquisition of Console Connect added the ability to provide users with uncontended SLA-based services across this global fabric, with the ability to provide highly-scalable and flexible, high-speed point-to-point connectivity on demand.
It is that capability which now makes building multi-cloud services that extend beyond a single campus a practical and commercial proposition. According to Paul Gampe, PCCW Global’s Chief Technology Officer, enterprises are now past the point of experimenting with development and testing workloads in the public cloud. Today they are moving mission-critical applications there for real, and that requires the use of direct connect services to make it reliable and affordable:
If you're a CIO and you're moving a mission-critical application that's used by 5000 staff, you've pretty quickly discovered that the public Internet may not be the appropriate form of connectivity to get to that mission critical application.
All the major public cloud service providers have their own direct connect services now, but that is only part of the issue when it comes to enterprise-level multi-cloud operations. There needs to be a way to orchestrate that level of connectivity in an environment in which individual enterprise architectures are, in effect, far-reaching ecosystems.
Gampe pitches Console Connect as the link between those enterprise ecosystems and PCCW, itself both a major carrier in its own right and a bridge between the eco-systems and all the cloud service providers. Its operations stretch from its own Layer 2 network, which includes high level inter-connects with all the other Layer 1 network providers, through to some 400 data centers around the world and direct inter-connects with all the major cloud service providers.
Like much of the telecoms industry, this is now the focus of increasing degrees of standardisation to ensure that global communications capabilities meet modern requirements, such as direct connection. To this end, PCCW Global is a significant contributor to the Metro Ethernet Forum with aims to promote ever-greater standardisation. According to Gampe, it has a suite of APIs and is developing a standard Lifecycle Management API for productivity.
Use cases for every cloud user
There are three primary use cases for Console Connect. The classic one is where a business or organization has a workload running in a far-flung data center and needs to move some very sensitive data to its disaster recovery site in, for example, London. This requires a 10gb bandwidth point-to-point pipe for 24 hours. Most telecoms, approached direct, would ask for a one-year contract as their minimum terms of business.
Effectively, PCCW Global buys bandwidth wholesale and sells it `retail’, says Gampe:
So what Console Connect does is provide a web interface and an API, which allows users to say, ‘I want to go from this port in this data center in Hong Kong, to that port in that data center in London for 24 hours, turn it on now’. And within 30 seconds, we automatically provision that - private, secure, quality-assured - and in many cases, we actually even own the submarine cable!
The company also has partnerships with all the leading telecoms carriers and they all now share routes and capacity to ensure that all users have availability on all routes.
The second use case is where one business or organization wishes to make a private direct connection with another without bothering with VPN tunnels, encryption or overlays – a guaranteed private, not-on-the-internet connection between that business and the third party. Here, Console Connect works in a similar fashion to enterprise networks by providing a connection that extends directly to their infrastructure. Gampe describes that as essentially “the world's largest Software Defined inter-connection platform - API by the hour.
The third use case plays to the increasing use of multi-cloud architectures as a core part of information and business management. It makes both business and operational sense to have workloads or logical data silos on the most cost and operationally effective cloud services, but that inevitably means moving them between different service providers. Gampe used AWS as an example this process in action:
One of the things that you discover, when using AWS Direct Connect, is that AWS provides a financial incentive to use direct connect. It can be 30% cheaper to move workloads in and out of AWS, and if you're moving Terabits of traffic, that's a lot of money. Direct connect is up to 30% cheaper. The alternative is the public Internet, if you're not using a direct connect product, at public internet prices. Not only is it cheaper, they get a guaranteed quality of service, and they can control the path from source-to-destination, which if they are dealing with GDPR or other regulatory compliance measures, is really important. As soon as you send packets over the public Internet, you literally have no control over how they move from source to destination.
The next step for the company is expected to be a move to address what is likely to be an inevitable problem of the near future, what NetSuite’s former CEO, Zach Nelson used to call the 'software hairball’. He was referring to was the growing mass of intra-company data connections, but with the growth of both the cloud and niche-market SaaS providers, this is starting to grow an even bigger ‘hairball’ where a single business process can now consist of multiple-source connections with multiple-source recipients and where multiple SaaS providers contribute to a single customer’s business process through multiple passes of data between them.
This model can be seen as a development of the growing ‘gig’ economy where the individual is replaced by the specialist, niche market service provider, delivering results as SaaS. For this to happen, it will require more than the point-to-point connection capability, no matter how fast users can set them up. It will also require the ability to provide and manage packages of orchestrated services that can be called up and implemented as and when required, ranging from the occasional workload through to continuous, dynamically changing orchestrations driven by AI-managed environments.
This same approach will also play well with the growing army of specialist channel partners that surround the major cloud service providers, where each can build their own orchestration packages to sit the services they can offer.
Seeing as the term cloud actually came from the telecoms business – where international connections became fully-automated and (just about) global, and the human operator-driven ‘book-an-international-call’ services were the order of the day – it is easy to assume that all such issues are now taken care of. But with the advent of cloud computing and, in particular, the growth of multi-cloud operations that can not only span the globe, but also need to do so while avoiding certain countries, it is an area which has rapidly started growing its own ‘hairball’ of connections that need management and implementation. So it is good to know that solutions are coming up as the need arises.