Connectivity has become mission critical. But how did that come, and what does it mean? One thing is certain: everything points to the cloud, because good connections are more important than ever. To understand the role of connectivity in the current success of cloud computing, it is necessary to look at the development of IT, business and society.
By Olav van Doorn, CEO and Rutger Bevaart, CTO
One of the major developments was the fall of the Berlin Wall, which resulted in the breakup of the Soviet Union. The world became less dangerous, and large parts of the globe and new markets were now more accessible. These markets had educated populations, but great material shortages. This allowed the Western business community to invest, and they were enthusiastic about doing so. With the iron curtain down, globalization was able to take a stronghold in the region.
Another catalyst was the liberalization of the telecom market in the second half of the 1990’s. State monopolies of European PTTs were lifted, AT&T was forced into the US to split up, and this led to many new businesses and numerous innovations.
New enterprises shot down the cumbersome, bureaucratic PTTs on all sides. And the lead time for new connections was reduced to days instead of months — from application through to commissioning. The mobile phone began its march.
A third milestone, related to the foregoing, was the public introduction of the Internet in the mid-nineties. Suddenly anyone could email and present themselves on the web. And they did. Everything would soon become web-based, web-centric or web-native.
At the end of the century came waves of the growth for the Internet, including a new and promising concept: Application Service Providing (ASP), enabling the online use of software applications and thus the precursor to SaaS (Software-as-a-Service). But not all the pieces of the puzzle fit correctly just yet.
For the user, ASP had too little added value compared to doing it yourself. The technique was not yet sufficiently developed to allow ASP to take great flight. In particular, the compounds were insufficient; Broadband was only at the beginning of its rise.
And then there was the burst of the Dot.com bubble — when the economy seemingly came to a halt. The orientation of more revenue from new opportunities offered internet to be shifted to the cost side. But after the bubble burst carriers had to make cuts, layoffs, and even file for bankruptcy.
One after the other carriers went bankrupt, and smart buyers were left with the assets. And without that financial stability, they couldn’t use the newly acquired networks for further innovations. The large-scale financial investment in network capacity was therefore written off in the bankruptcy proceedings.
Cloud and Virtualization
But soon, it was soon business as usual. Internet became a commonplace, and broadband connections were a commodity. Virtualization made payment to the possible server hardware more effective, and ASP suddenly became interesting and made the ‘Cloud’ a formidable start.
The term Cloud computing had already been introduced in 1996 by a group of engineers at Compaq Computer. And by August 2006, the concept of the cloud was presented to a wider audience for the first time by Eric Schmidt (Google CEO) at a conference.
Since the end of the last decade the Cloud Computing concept has come a long way, including here in the Netherlands. The initial period was still marked by extensive discussions on definitions and whether or not it was a real innovation. It was frequently suggested that The Cloud was just, “old wine in new bottles”. But, we are past this phase.
First Cloud, and Cloud Only today, are the primary IT strategies and embraced in the enterprise world. Fears over the unknown and concern for safety have been remove.
The Cloud has also become essential for the providers, and they deliver cloud services in a variety of ways: from hosting and data center world, to the carriers, to the system integrators and pure players, and also the large public cloud providers. Each of these categories has its history and origin.
The importance of virtualization for the development of cloud can not be overstated. Thanks to virtualization, hosts could put their expensive server farms to much better — an outcome in lean times after the burst of the Internet bubble. Virtualization made it possible for companies to transfer their applications without downtime to other locations, such as a data center or hosting environment from third parties.
The genesis of hosts is extremely fascinating. Most started in the attic with the registration of domain names and hosting websites, later moving on to storage. And with the growing importance of the internet, hosts saw an opportunity to make their jobs professional, so they started standardizing and automating their work extensively. Often their service package was extended with data center and colocation.
In the early years of cloud, it seemed for a while that the local players would be wiped out by the big American parties. It was clear that consolidation going on. Plus, there was competition from an unexpected source: the traditional system integrators who were working to make their antiquated business model of hourly billing and projects close to the modern era of subscriptions, services and functionality ‘off the wall’. That means they also provided hosting and SaaS.
The system integrators have had to make some big transitions. Along with providing hardware and software, prehistoric IT was extended for a while longer with value added services. Do you remember: VAR (Value Added Reseller)? And soon after came the development and integration of systems, like posting and body shopping. This integration was before the advent of SOA, and web services and APIs were then often made unnecessarily complicated. The advent of The Cloud has largely eliminated the need for the development for new systems. However, the demand for “unburdening” through managed services is greater than ever. And those who think that the world has become less complicated with the advent of The Cloud is wrong. There is just an added layer of complexity, which cloud systems have integrated and managed with each other along with existing IT. There is the added value of the system integrator new style: hosting is part of this.
Of course, there have been carriers of interest alongside the strong growth of The cloud. Cloud use by international enterprises require high quality connections in some of the most exotic places on earth. Carriers have been used from the time of the old phone, to branches internationally. And in keeping pace with the development of the Internet, carriers provide carrier leased lines. They have made huge investments in recent decades to accelerate and broaden and their compounds — often under the most challenging circumstances.
It will come as no surprise then that the carriers want to move up the food chain, and they are going to do so by focusing on data center services. Carrier hotels bring connectivity and housing under one roof. Remarkably, the promise of the carrier hotels- at least for the provider – was not fulfilled. And a number of carriers have already returned from their datacenter adventures — but perhaps their input was not good. For cloud and data center services to sell you should go to the CIO, and not the traditional party of the carrier: the telecom manager. In addition, they were too late. The hosts already had everything in order.
The data centers have established themselves firmly in the heart of the cloud infrastructure. Thanks to the creation of Cloud Exchanges, they’ve become hubs for access to various cloud providers, and companies can now connect from the datacenter to the cloud provider of choice. This allows them to stay where they are rather than having everyone come to the cloud providers. This development also fits the needs of the user to hybrid cloud environments to public clouds, as it can remain in the private cloud as it should. The consolidation of providers that is currently underway reinforces the power of the data center provider.
The high expectations of the public cloud providers have yet to be fully realized. The acceptance of Amazon, Google for Work, Office 365 is steady but relatively slow. Presumably, we saw the same in the early years of the Internet: the show goes slower than predicted, but once it gets going the impact is much greater than anticipated.
Another expectation that has yet to be fulfilled is all data traffic running over the internet. And now the question is whether this will ever happen. Internet is robust and reliable, but it does not come with guarantees and SLAs. Especially when it comes to business critical or extremely privacy-sensitive traffic, companies will continue to go through their own connections. However, the quality of the internet is getting better. Synchronous communications such as telephone and video conferencing over the Internet is now much better.
What does all this mean for connectivity? It should be clear that the current wave of digitization only reinforced the key role of the data connections. And a new boom arrives with its own network infrastructure and protocols: Internet of Things represents additional demands on communication. Again, a custom approach is necessary. Some IoT applications are small amounts of data at regular intervals, for example: the measure of water in houses. Others, such as the self-driving car, accompanied by large volumes of real-time information should be safely sent and processed. The data communication manager will also keep in mind the control needed over this kind of communication.
Due to the increasing complexity of networks, from a risk management perspective and in line with the trend towards outsourcing, companies are already inclined to take off all of their communications from one party (one-stop-shop). And they are right, only they often make the wrong choice. They choose a major carrier with the expectation that they are the easiest and best — which we now know is a miscalculation.
Proprietary or neutral?
In connectivity, there are broadly speaking two kinds: carrier-driven (supply-driven) and carrier-neutral (demand-driven). Choosing the one-stop shop carrier, which has a carrier-driven variant, means choosing the offer of that carrier. The carrier then puts in place its own infrastructure and ensures that partners are engaged in the regions which the carrier itself does not cover. This approach has three disadvantages: the infrastructure of the carrier is often not fully up to date, the carrier will offer a one-size-fits-all solution when it is better to look at an application and from the business from the customer’s needs, and finally SMEs do not always get the attention they need because the major carriers are more focused on their customers in the large enterprise segment.
Some carriers implement the one-stop-shop concept but also go one step further by providing connectivity data center services from their so-called carrier hotel. But the disadvantages to this is the possible exploitation of a data center that is not the core business of the carrier. And there is also a large degree of vendor lock-in in the relationship. We have already noted earlier that this trend is past its peak.
In contrast, the carrier-neutral approach allows for much more freedom and flexibility in the choice of the solutions. Competition between the various carriers is fierce, and that in turn pushes prices down and enhances innovation — which works directly for customer benefit.
Also in the carrier-neutral approach, it is possible to use a one-stop-shop provider; this will be based on the business requirements of the customer design and management of a customized infrastructure. That is no easy task, because the connectivity landscape is extremely dynamic in terms of supply and in technology. High-quality specialty is needed to fulfill this one-stop-shop role well. A good solution contains the right mix of cross-connect (direct connect via MPLS), and Internet and Cloud Exchange. And in the end it has to all be centrally managed and supported, with easy self-service capabilities for customers.
One-size-fits-all does not work
What are the variables? Independent of the selected compound technology, the choice of connectivity is determined by four parameters: price, bandwidth, latency (delay) and availability. Each situation requires a specific consideration of the optimal mix. And each choice therefore has an impact on the selection of the connection technology, and the network providers. One company might need a lot of bandwidth at a low price, and another company may need the lowest latency and highest availability, and is prepared to spend a little more.
It’s necessary map all these needs, and the user’s application landscape. Because each application category must be determined from risk management and cost perspective to figure out which connectivity is needed.
Often though, companies choose a one-size-fits-all approach that seems simple and clear. But in our firm conviction, and from a cost effective point of view and risk management policy, it’s much better to regulate the bandwidth required for each application. It is simply not necessary to align the company-wide connectivity to the peak load of a single business-critical application. In addition, the carrier-neutral approach, a one-stop-shop, is also simple and easy for the customer.
Connect the Cloud
In a cloud environment, the chain is only as strong as its weakest link — and all too often the weakest link is the connection. It is still possible for exerting influence on the outsourcing partner, but cloud computing is different.
With service providers, you have one offer to accept. Then you also become dependent on their connectivity. Experience shows that cloud providers are very focused on their internal systems and their technology, providing high SLAs. This means the connectivity remains underexposed or is not included in the SLA. In other words, you get “guaranteed access to the front door.” Moreover, we see that some innovative cloud providers have gotten an eye for the importance of connectivity for the quality of the service provided. They offer next alternative internet connection methods with the required redundancy, and provide guarantees. But, the concept of carrier-neutral Cloud Exchange in the datacenter gives a strong boost to the quality connection to the cloud provider.
Connectivity in today’s cloud economy is crucial. A customized connectivity policy should be aligned with the business strategy, and contribute to good risk management. A smart approach is to save money, increase the overall availability of IT services, to increase the flexibility of the company and to strengthen its competitive position.