Managed Services

Enterprises Not Rushing to Embrace Cloud

Enterprise IT departments have not embraced cloud services with the fervor once expected, say executives at two major telecom providers who are delivering cloud as well. The reasons are varied, but many are the same today as they were three years ago: IT execs fear moving their critical apps to the cloud means losing control, facing greater security risks and being locked into one provider's cloud.

So while cloud services are growing, for them to succeed as expected, telecom operators have to work harder to help enterprises through the transition and have to improve what their clouds deliver to make it more closely match enterprise requirements.

Doug Junkins, CTO of NTT America Inc. , spoke recently at a conference in Singapore and says he found enterprise executives were mostly raising the same familiar issues. "We are still trying to debunk the same common misconceptions about security, resource control and migration," he says. "There is a level of conservativeness in a lot of enterprise IT executives -- they are not willing to make what they see as wholesale changes."

A recent Verizon Terremark blog on debunking myths about cloud addressed the common issues with language similar to that of cloud conferences held in 2011, an indication that things haven't moved that far that fast. Verizon Terremark's sales of cloud services are increasing, says John Considine, CTO, but that's because business units within enterprises are going out to buy their own cloud-based resources, not because cloud services are being warmly embraced by the IT departments. (See 7 Truths About Cloud Computing.)

"We have taken a careful look at who is buying cloud services from us and in the enterprises that are adopting cloud computing, a lot of the work is coming directly from the business units, from the chief marketing officer, the R&D organizations -- non-traditional IT buyers. In many cases, we are pulling the IT folks along with us," he says.

Considine believes IT executives' fears about cloud services are being fed by hardware and software vendors who stand to lose sales if enterprise applications move to the cloud. "Server vendors, storage array vendors, network vendors -- they are not really interested in having cloud do well," he says. "They will help anyone who wants to, not use cloud. There are a lot of people trying to subdue cloud -- but even in the face of that kind of background drag on the system, cloud continues to grow."

Both Junkins and Considine believe enterprises will be won over by the long-term benefits of cloud-delivered services and apps -- lower capex and opex, greater flexibility in responding to market demands, and even, perhaps ironically, faster disaster recovery. But they say it is going to take more effort on the part of service providers and continuous improvement both in the cloud services themselves and in how the transition to cloud is managed.

"We are taking an educational approach," says Junkins. "We need to be able to demonstrate these benefits clearly to our customers. Any time your servers are connected to the Internet, there will be some level of risk and cloud isn't any different on that score. But you can use data duplication and move service from one location, so that if there is a disaster in California, for example, and the data has been backed up, then we can have you up and running more quickly on servers available in Virginia."

Considine says his team is constantly looking for success stories, where enterprises are using cloud to a competitive advantage, and he expects to see cloud continue to gain momentum in this way as well.

But the other piece of it is acknowledging that cloud services still need to mature, he adds. Verizon has been actively promoting what it calls the "cloud of tomorrow" via blogs such as this one, trying to drive home the difference between the service it offers and what is more commonly thought of as commodity clouds such as Amazon Web Services and others. Verizon also is talking about what needs to happen next for cloud services to succeed across a broad spectrum of enterprise applications, including some mission-critical ones.

More commodity cloud deployments require enterprises to plan for the potential failure of any component at any time, Considine says. That may work for applications that are stateless and can easily horizontally scale, but many critical enterprise apps don't fit that profile and would have to be substantially re-written -- a process for which enterprise IT folks don't always have the talent or the tools, he adds. In addition, all the effort and expense enterprises have spent around performance management goes away with a commodity cloud service, and the enterprise is forced to accept technology choices made by its cloud vendor, which leads to the loss of control and insecurity many enterprises fear.

"What we are asking is, are these attributes of cloud computing or the current implementations?" Considine says. "By doing cloud better, more workloads can go into the cloud, enterprises can find it easier to take advantage of it and you actually reduce the aspects of lock-in cost and the burden of making the transition and you end up with an overall better solution."

That's the solution-sell more telecom cloud providers are likely to be taking.

— Carol Wilson, Editor-at-Large, Light Reading

Page 1 / 2   >   >>
brookseven 8/26/2013 | 2:38:58 PM
Re: Self-preservation  

The fundamental problem is that we define all things "The Cloud".  I tried to break that into several categories earlier and yes I agree for a percentage of applications public cloud offerings will not be used.  On the other hand...SalesForce.  :)  Not much more private than customer information and price lists and yet SalesForce is going gang busters.

My point all along is that different solutions will work for different folks.  On top of that between time and cost nobody can move everything at once.  An enterprise might move A major application once every few years.


sam masud 8/26/2013 | 2:30:57 PM
Re: Self-preservation Thanks, and I see your point. But basically there are two types of customers--the mass market and businesses. Public clouds would work for the mass market, but private or hybrid clouds might be what businesses would prefer.
brookseven 8/26/2013 | 2:02:17 PM
Re: Self-preservation The whole point of "The Cloud" is internet connected virtual environment.  The whole thing started around using public, shared resources in a virtual manner.

How does that become "Private"?  At that point, people are just using IP networking to 3rd party data centers.


sam masud 8/26/2013 | 1:35:42 PM
Re: Self-preservation I'm curious why you hate the term Private Clouds....
mark-r 8/26/2013 | 1:14:45 PM
Re: Self-preservation Perhaps the key addressing any valid (ie not self-preservation type) obstacles for enterprise cloud adoption is to allow (but not require) the enterprise app developers to have, as much as desired, in-depth control over their application deployment and execution environment in the cloud. An open platform as a service, with customizable development environment as well access to 'bare metal' server hardware would -- while providing the cloud benefits of virtually limitless, elastic capacity, horizontal scalability, everything on-line etc. -- enable full developer-controllability of all relevant aspects of running their applications.

Of course it also matters what the provider hosted server hardware is, in particular for high performance apps: for instance many of the ultra-low latency trading, risk calculus and some specialized work flows such as technical/scientific simulations, genomics algorithms etc. get their performance from specialized hardware such as GPUs and custom logic (FPGA) processors.

It is a challenge, but the progressive cloud providers will need to enable the application owners to control the performance impacting aspects of running critical applications in the cloud. The key to success is a platform (as a service) that makes it more productive to develop and deploy horizontally scalable, high throughput, low latency apps in the cloud than on the enterprise's own IT facilities.
telecom007 8/26/2013 | 7:21:10 AM
Re: Self-preservation If there is any "Massive" movement of the Enterprise space to the Cloud it's more on the likes as someone said "Private Cloud:".  Just look at the continued explosive growth of Data Centers.  Rack space in Data Fortresses have become no-brainers for Disaster Recovery or Better security than keeping all Enterprise-owned hardware on Prem.

For CRM services like Salesforce, yes One Button Bob might apply but you don't see that same ease/smoothness with Exchange or just about anything Microsoft whether it be the OS or Files.

As the article stated the Turf protection by Hardware vendors is a reality as I've also seen that in the same vein by Hardware vendors for Premise-based PBXs vs Hosted PBXs.  

Eventually the Vendor Turf protection will lose out over time but this is a longer play than what some were predicting.

Carol Wilson 8/24/2013 | 7:33:02 AM
Re: Resistance to change Phil, restrictions on where data can reside is definitely having an impact on cloud services, especially in Europe. 

And brookseven, I like the phrase One Button Bob and I plan to re-use i liberally. 
Phil_Britt 8/23/2013 | 6:33:07 PM
Resistance to change Even though people certainly will resist change, even some of the most resistant (financial services companies) are moving to the cloud now. But some will use only hybrid or private cloud models, especially for critical data.

The financial services providers, health care providers and others need to know where data resides, and there are limitations on the information that can reside on offshore servers. So some may never adopt full-blown cloud infrastructures, but the trend is certainly there.

brookseven 8/23/2013 | 4:43:28 PM
Re: Self-preservation Carol,

Actually, I would argue it is the normal round peg problem with technologies.  There are places that various Cloud solutions work best, but it will never be 100%.  So, will the cloud take over everything?  No.

Secondly, any Enterprise can only change so many things at a time.  It is easier to adopt new technologies where they are seamless OR are greenfield.  Walking into a company and say...lets retire your Exchange Infrastructure...is a work of migration.  Migration ADDS work (and thus people/time) during the migration.  What I rarely see is what we used to call "One Button Bob" migration tools.  Press a button and the tool reads the data from the old tool and pre-populates the new one.

The reason this bit is often ignored is that (especially at the beginning) technologist tend to be evangelicals.  I am better/faster/cheaper so buy me!  Well, what if you are saving $100/year on a $1M budget?  You are better but its a LOW priority.

I remember talking at a Forum where I was next to Nan Chen talking about Carrier Ethernet.  I pointed out that Enterprises could really use the firing of the guys who know about T-carrier services and that was an awesome OPEX savings.  Nan (to paraphrase) said that OPEX savings was not about firing people.  But of course, that is exactly what it is about if it is going to be meaningful.  Most new technologies don't really think through the personnel restructuring.

In fact the way I would put it to a CIO is that he or she could redeploy the budget on higher value items and put rote services to a 3rd party.


Carol Wilson 8/23/2013 | 12:26:01 PM
Re: Self-preservation I think the points you raise are things with which my sources - Doug Junkins and John Considine -- would agree. 

People are resistent to change, which is why they want the new CRM to behave just like the old one. There are persistent concerns about turning over information and yes, as I noted yesterday, outages are all too common. 

That's why I think it's interesting that companies like NTT and Verizon are trying to evolve how they deliver cloud services. They may be facing impossible odds, as you seem to be saying, but I think they at least recognize the nature of the challenge. 
Page 1 / 2   >   >>
Sign In