opinion


A cloudy outlook

cloudhead

The technological ‘cloud’ is as vast and nebulous a term as its meteorological namesake. It means many things to many people and presently rides the crest of the hype cycle, where like many technologies before it, it is often hijacked for marketing purposes. To this end the cloud is easily confused with utility computing, grid computing, virtualisation, or even networking. In truth it is all these things and more, with the key feature being that resources, whether they are applications, platforms or infrastructure, are delivered as a service, from a non specific origin (the cloud) and are billed as a utility, like electricity.

Despite the hype that surrounds it, cloud computing, in its broadest sense, is not a new concept. In the wake of the dot com bust, traditional internet and IT firms such as Amazon, Google and IBM, which found themselves with a dearth of computing capacity, began turning the internal efficiencies deployed across their own data centres, into a commercial product. A new business model emerged. By passing these savings on to their customers, through the introduction of pay-per-use services, the first cloud providers also found a new way to cover the costs of their own infrastructure usage.

It’s an economic no-brainer. All the reasons that cloud services are interesting to an operator’s customers are all the same reasons they are interesting to the operator. Telcos are huge consumers of IT resources, so consolidating the approach to internal and external demand should give better savings.

According to Camille Mendler, analyst head of enterprise solutions at Informa Telecoms & Media, most of the focus on cloud enablement to date has been on setting up telcos to sell on cloud services to their customers. “But this is putting the cart before the horse,” Mendler says. “Telcos need to clean up and streamline their own operations before using them to sell on these cloud services.”

Jose Luis Gamo, global cloud services director at Telefónica, agrees: “The cloud space is one in which scale is important. Besides, we can leverage our global footprint to distribute the data centres geographically, tackling issues that range from the regulatory to the functional,” he says. If RIM had been able to store data locally, for example, it might not have had to endure its recent difficulties in India and the Middle East.

HP, a traditional cloud enabler, has reaped the benefits of the cloud in its own organisation, offering a useful example to carriers. In the summer of 2010, HP invested $1bn in its own commercial data centres, management platforms, networks, tools and applications to create a more scalable, modernised and automated IT infrastructure. HP as a business already runs on its own offerings but, as a result of this recent investment, and the ensuing productivity gains and automation, the firm is in the process of eliminating roughly 9,000 positions over a multiyear period and expects to generate annualised net savings of between $500m and $700m.

These savings will be passed on to HP’s customers, many of which are telcos. Tim Marsden, director of communications and media solutions at HP, says: “It’s hard to pinpoint an exact number, but scaling will definitively lead to savings.” Bear in mind of course that operators are not always looking to save only on cost. “In cases where we had cloud OSS discussions, a key driver for the operator was not necessarily the direct operational costs but leveraging standardisation. A SaaS (Software as a Service) solution will pre-empt business architecture decisions and—assuming that the SaaS platform will converge towards a best-practice approach—the operator will reduce engineering costs. In this way, SaaS helps mitigate the ‘Not Invented Here’ syndrome,” Marsden says.

Cloud busting

What exactly is a cloud service? Broadly speaking, cloud services fall into three categories:

IaaS (Infrastructure as a Service)

This encapsulates pretty much any kind of hardware such as servers, racks, or data centres both in the private cloud and in the virtual or public cloud. Organisations could dynamically add more web servers or storage capacity as an example. Sometimes called Computing as a Service (CaaS). Amazon Web Services fall into this category.

PaaS (Platform as a Service)
Third party providers or operators can provide access to public toolkits allowing organisations to develop their own custom software applications. These platforms are virtualised across servers and datacentres, negating the need to invest in hardware and other provisioning capabilities. Examples are Google’s App Engine or Microsoft Azure.

SaaS (Software as a Service)

This is software on demand. A cloud services provider switches the traditional licensing model for software to a pay-as-you-go, allowing users to switch resources on and off when needed. Along with rapid scalability, users benefit from a single installed version of the service that is also managed by the provider. Salesforce is on example.


Along with standardisation, one of the defining features of a cloud deployment is elasticity—the ability to shrink and grow resources almost at the push of a button. “From two weeks to two clicks,” is a phrase doing the rounds on the tech circuit and the most obvious areas where telcos can gain efficiencies is in their own computing power and infrastructure. If a company uses hardware—servers, racks, or infrastructure of any kind—it must involve itself in time-consuming requisition activities.

But cloud computing drastically reduces the time it takes to order and build these things, as well as the cost, because you don’t need staff to set up those servers, racks, cabling and other bits and pieces. It’s the same business concept that the operators are selling to their customers. The “as a service” model is becoming a real alternative to in house, as it’s much cheaper and quicker, pay per use rather than capital expenditure, and it reduces the need for IT resources. “Leveraging a multi-tenanted platform offers much better economies of scale than you could get by yourself,” says HP’s Marsden. “Larger telco groups have expanded through acquisition and each operation works independently of the others, often with their own legacy platforms in place. Now these operators are trying to consolidate into one private cloud, so they have one central version of an application or platform, and get rid of legacy technology.”

Orange is one such operator, with multiple operations around the world. Stephan Hadinger, chief architect of cloud computing and head of the Orange API/Telco 2.0 business unit reveals that Orange kicked off its internal consolidation four years ago. The company started its transformation with a consolidation of software platforms and mass virtualisation of servers having partnered with specialist firm VMWare and escalated the project towards a true cloud from there.

“In 2006 I was working on the Orange IPTV platform, where I managed the growth from 30,000 customers to one million customers. So you can imagine the scalability issues,” says Hadinger. “This was the first project where we really introduced server virtualisation and VMWare products, giving us additional flexibility and agility. The main thing I was looking for was the ability to clone servers quickly and easily without needing to reinstall the server, which can be a very costly process. So we started the collaboration with VMWare.

“People thought we were crazy putting virtualisation on production services and more so because it was for IPTV which is almost a real time service. We had some issues but it worked very well and today we have 14,000 virtual machines. Orange can really benefit from the scale effect in being able to leverage some big infrastructure.”

According to Hadinger, Orange has reduced the time it takes to deploy services from several months to several days. Orange Beta, as it is known, is a cloud-based platform designed to get applications or services up and running in production within one week. “It’s the first time we put an objective and KPI on time to deploy metrics,” he says, “So you don’t have to spend a lot of money industrialising a solution if you are not sure it will be a success.”

Hadinger manages what he calls “agile hosting” for Orange internally as well as externally. This is not only the infrastructure but also the 24/7 operation teams that manage all the applications and virtual machines. For internal needs Hadinger revealed that Orange has over 600 virtual machines running over 35 projects and offers availability of 99.93 per cent from end-to-end, including connectivity and operation of applications.

“We aim to first deploy this federation approach within the company,” he says. “We are already sharing a lot. We are able to define platforms across the group and each country is using the exact same components. We are not yet at the point where you can choose from your web browser to use resources from Switzerland or wherever, but this is the way it has been designed and it will have this level of federation in the future,” Hadinger says. “It’s a good way to test federation and show customers how we will use the same technology we are selling them within our own company.”

There are benefits for newer carriers as well as established players with a mess of legacy systems. HP’s Marsden says that a number of telcos the company is working with are rolling out cloud based operations for their own business, with some going as far as deploying, “their accounting, network management, everything” in the cloud.

French operator SFR, one of HP’s flagship cloud customers, was looking at how to leverage an internal cloud for its own operations when it first recognised a big opportunity to cater to its existing customer base by offering backup, hosting, conferencing, and Microsoft applications among others. To some, this is known as a ‘hybrid cloud,’ and represents the adoption of a holistic view, whereby a company uses the cloud for its own internal efficiencies, then leverages that scale to pay for its enhancements and make more money out of it.

Cloud technology adoption internally typically originates, as Orange’s Hadinger acknowledges, from the IT department, which wants to show efficiencies to the CFO. External facing cloud initiatives on the other hand tend to originate from the marketing department. But in many cases, operators like Orange, SFR and Telefónica, are now moving to unify the two strategies.

In a sense, this transformation changes the position of the telco in the market, from a pipe provider to an integrator and application provider. “Telcos have been striving to become integrators for some years but only a few have been successful,” says Informa’s Camille Mendler. “Enterprises look more to integrators, as there’s more involved in the integration work than there is running the services, but once you’ve done the migration then it becomes all-out war.”

A traditional cloud provider such as Amazon, for example, might offer a contract that covers the uptime of the data centre but not the connectivity. An operator, on the other hand, can deliver an end-to-end agreement covering the service, the application, the integration, the connectivity and more. “And that’s the fundamental difference,” says Mendler. “Cloud has typically been talked about as an IT thing but it’s 100 per cent dependent on network,” she says.

This is why partnerships with the cloud enablers: Amazon, HP, IBM, Microsoft, Google, VMWare, Cisco, NEC, to name but a handful, are so important—especially because of the opportunities they offer for operators to differentiate. Informa has expressed concern that the telco community is coming out with fairly generic cloud offerings, from basic computing power to storage. “Most operators are just thinking about generic IT functions,” says Mendler. “Telcos are worrying about the margin falling out of connectivity after they’ve sold all the IP VPNs and voice and data services they can to customers, but they are all selling the same cloud service,” she says.

“Telcos are not developing new applications. They are using SaaS as a loss leader, to get their foot in the door to sell IaaS (infrastructure as a service), because that’s where their asset ownership is. But if there is no differentiation and it just becomes a commodity service, they are repeating the same mistake as before. It’s a move from dumb pipe to dumb cloud and then it will just become a price war,” Mendler adds. “The key is to start selling vertical cloud solutions, which have very specific security and management requirements associated. But to do that, telcos need more credibility.”

This is where the partnerships come in. “Partnerships are critical for us,” says Telefónica’s Gamo. “One has to be conscious of one’s position in the ecosystem and have great ambition, but never leave realism: We are a service provider and not a technology designer or manufacturer, so we need best-ofbreed technology to build our services. And we need this technology to be state-of-the-art, cost-effective and flexible for its integration with our operations support and business support systems,” he says. “In these partnerships we are looking for this and for a joint go-to-market, to improve the effectiveness of our commercial action. As for the value that we add to our partners, I think that it lies mainly on our experience in providing mission-critical services with stringent SLAs, our channels and our brand.”

Telefónica has a strategic partnership with NEC for SaaS and virtual desktop infrastructure (VDI) and, while Gamo recognises the importance of this partnership he is clear that this is not the only one, noting the operator’s partnerships with VMWare, Cisco and EMC for IaaS.

“What are the basic needs for large scale deployment?” asks Manuel Gallo, head of business development for cloud computing at NEC. “You need network infrastructure and bandwidth, and then you need users—sometimes consumers or sometimes businesses.

So we looked at who was best positioned in the market to provide this and decided that for a large scale, widely applicable cloud, the carriers are the ones making it possible today. They have high amounts of bandwidth, both fixed and wireless and big data centre infrastructures. They own the network and they also own the customer relationship. So we started strategically partnering with carriers for cloud business,” Gallo says.

“The average time a carrier needs to implement a third party application into their systems—such as a billing system, is 12-18 months. So we reduce that time by delivering the software as a service and as well as sharing the risk we share the revenue,” he says. “We have created a partnership model for three parties: the carrier who owns the network bandwidth and the users; the aggregators and integrators; and the software vendors which deliver their products to the users. We sell not just technology, but also relationship and contractual issues. “An end user or a carrier isn’t ready to establish relations with 20 to 50 ISVs (independent software vendors), but we are.

We do the technical integration and all the payment schedule and billing for the ISV,” he adds. “Then the carrier has a pre-existing relationship with the customer and they can just add this service on to their existing bill.”

Billing is a big consideration in the world of cloud services and, again, it works both ways. Operators can offer any combination of SaaS, PaaS, and IaaS, and support tens of third-party providers that want to install an application and offer it via the telco’s cloud. But the end user doesn’t want a relationship with each and every application provider.

To return to the “two weeks to two clicks” idiom, they want to be able to provision more computing power or another service at the click of a button and then have that charge added onto their existing bill for the duration they use the resource.

But as Mohammed Sha, global head of market insights at billing provider Convergys points out, operators can also leverage cloudbased offerings to make their own business more efficient and cost effective, especially those with spread out operations and several billing systems running in several different locations.

“Consistency in an operator group’s billing platform is not possible in this world. But the telco could externalise its entre billing infrastructure for all subsidiaries, with a centralised billing engine in France, for example, which every subsidiary hooks into. This way, you get consistency across all subsidiaries and benefit from time efficiency as well as cost. Because it’s only a single instance as opposed to multiple different areas of procurement, you can leverage this private cloud to benefit your own business operations,” Sha says.

Convergys provides a hosted and managed billing service in its data centres. Then whenever a new operator needs a billing platform they just hook up to the data centre. When installed across the whole business, the provider also takes over the support and maintenance of the platform. “This removes the huge cost and hassle of dealing with several providers and takes away the entire headache of procuring a billing system and buying in infrastructure,” Sha says. “Every time there is an upgrade or new feature we do it centrally and the operator gets the benefit from it across the board.

“Of course, operators need to be able to bill for services at wholesale and retail level too, like in the MVNO business or M2M. We can use a wholesale billing system to monetise third party services on the operator’s network. The end user sees no difference, and that also means the enablers don’t need their own billing infrastructure either.“

So really, the cloud model is just the next evolution of managed services and it offers just as many benefits for large, disparate operator groups as it does for small, rapidly expanding start ups, and their customers. Enterprises are beginning to adopt cloud computing as a means of offsetting capex with a more predictable opex spend, and Orange’s Hadinger believes that the model “is a virtuous circle”. Before the cloud came along it was very rare that a project could actually give back resources. Typically, it’s very difficult to put incentives on a project to reduce resources, but because cloud computing works on a pay-as-you-go model, the users of the technology have more transparency and have more of an idea how much a project really costs.

It’s also a natural business shift for the operators, which have an existing relationship with potential cloud customers. As HP’s Marsden notes, there is room for someone to enter the value chain and sit between the cloud provider and the end customer and add real value.

It sounds like a win-win and in many ways it is, because the cloud represents a true shift in business model, turning the one way supply and demand paradigm into a circular value chain. Most often, the focus in this industry is always on the latest and greatest technology and its promise of greater efficiencies, yet in this case, the technology itself is mature and tried and tested. The cloud is a revolution for its users, but it doesn’t matter to anyone whether it’s a technical revolution or not.

Who’s ahead in the cloud?
SFR
French operator SFR tapped HP to build it a complete cloud services platform to enable it to offer IaaS to local companies. Working on the hybrid cloud model – SFR is using the platform on its private, internal infrastructure to gain efficiencies in its own business as well as selling the same service publicly – the HPbased offering is hosted within SFR data centres. The HP platform will also allow SFR to evolve its offerings beyond IaaS to include additional cloudbased services, such as communications as a service (CaaS) and PaaS.

Orange
Orange Business Services has partnered with Cisco, EMC and VMware, under a business alliance called Flexible 4 Business. Orange delivers the four types of pay-per-use managed cloud services from the vendors, with a services portfolio composed of IaaS, including private cloud and back-up, and SaaS, including security and unified communications. As the service provider Orange will commit to tiered end-to-end service level agreements (SLAs).

Telefónica
Telefónica is working with NEC to offer virtual desktop services, and with Cisco, EMC and VMware to deliver utility computing and a cloudbased application marketplace called Aplicateca, which offers apps and services from ISVs.

Telecom Italia
Italy’s leading carrier launched a portfolio of cloud services at the tail end of 2010, targeting large and medium-sized companies and government entities with IaaS, PaaS and SaaS offerings. The initiative featured eight new data centres, a national technical assistance centre, and a security operation centre, which gives the company “significant economies of scale, which it passes on fully to its customers.”

Verizon
US carrier units Verizon Business has offerings catering to SMEs based on Google Apps and Microsoft Online Services, delivering commonly used business applications via the cloud. The company also offers computing as a service based on a platform from VMWare.

AT&T
AT&T Business Services offers a portfolio of hosted cloud solutions under the AT&T Enterprise Hosting Services banner. Hosted security, storage, content distribution, load balancing, utility computing and virtual services are all available and powered by VMWare.

NTT
NTT Communications, the long distance arm of Japan’s NTT, offers colocation, managed hosting, and VMWare-based IaaS to multinational corporations. It also offers its own content delivery network.

KT
South Korea’s incumbent carrier, KT, tapped up Citrix and Microsoft at the end of 2010 to build it a cloud platform geared towards offering virtual desktop services from the first quarter of 2011 as well as business tools under the Office 365 banner. The carrier also acquired NexR, a specialist firm in cloud dispersion storage and processing technologies.

Amazon
Although not an operator, Amazon is recognised as an innovator in the cloud IaaS space and is a pioneer of highly automated, inexpensive, commodity infrastructure, bought without any commitment to a contract. It also offers cloud storage, cloud-based content delivery and a number of PaaS-like services.


One to watch
Julian Turner is head of business development at cloud services enabler Private Planet. The start up offers an operator branded personal cloud platform that can be sold on to subscribers, with an application that works across the iPhone, Mac OSX, Symbian, Windows, Linux, BlackBerry, Android and Meego platforms. Features include personal data storage, cloud-based communications and applications running in an open personal server environment.

“From an operator perspective every time someone uses Apple or Google the operators are not getting any share of the revenue, they’re getting a bit of data revenue perhaps but they’re not getting a share of the real revenue,” Turner says. “The operators need to be selling a bolt on so users get the first bit free, a bit of storage or something, then the operator adds its own value and gets the lion’s share of the subscriber dollars.”



2 comments

  1. Raoul Butler 10/02/2011 @ 1:38 am

    Cloud is the user pays system… just what we need another grid to plug into, slavery by any other name.

    Why do I object… user pays sounds reasonable except when it becomes a commodity and when it fails… Look at all communication companies, they all have miserable support systems, they all work on user pay, if your a small fry, you get small fry support. It’s another way large companies cheat the user.

  2. Kais Belgaied 28/03/2011 @ 8:16 pm

    Raoul Butler, your concern about the quality of support for small consumers is valid, during a transition period only though.
    Using the electric utilities analogy, it took a couple of decades between the advent of the first public utility companies in the early 1900’s, and the conversion of most businesses to buy energy off the grid instead of operating their own little generators on site. Competition, economy of scale and technological progress raised the dependability requirments while putting pressure on the prices.

Leave a comment

Your email address will not be published. Required fields are marked *

Polls

What is your name?

Loading ... Loading ...