The Race to Zero


This is my 500th blog entry, which means I have written several books worth of words. When I started the blog I feared I might run out of ideas in a few months, but our industry has become so dynamic that I am regularly handed more topics than there are days in the week.  

The cloud industry is often characterized by what is being called the race to zero. This is the phenomenon of ever-dropping prices for data storage. The race towards cheaper prices has always been driven by Amazon through repeatedly reducing their cloud storage prices. Every time they reduce the price of their AWS storage services, the other big cloud companies like Microsoft and Google always go along.

There are numerous reasons for the price drops, all having to do with improved computer technology. Memory storage devices have dropped in price regularly, while at the same time a number of new storage technologies are being used. The large cloud companies have moved to more efficient large data centers to gain economy of scale. And lately the large companies are all designing their own servers to be faster and more energy efficient, since energy prices for cooling are one of the largest costs of running a data center.

I remember in the late 90s looking to back up my company LAN offsite for the peace of mind of having a backup of our company records. At that time our data consisted of Word, Excel, PowerPoint, and Outlook files for around twenty people, and I’m sure that wasn’t more than a a dozen gigabytes of total data. I got a quote for $2,000 per month, which consisted of setting up a shadow server that would mimic everything done on my server, backed up once per day. At the time I found that was too expensive, so we stayed with using daily cassette back-up.

Let’s compare that number to today. There are now numerous web services that give away a free terabyte of storage. I can get a free terabyte from Flickr to back up photos. I can get the same thing from Oosah that allows me to store any kind of media and not just pictures. I can get a huge amount of free storage from companies like Dropbox to store and transmit almost anything. And the Chinese company Tencent is offering up to 10 terabytes of storage for free.

It’s hard for somebody who doesn’t work in a data center to understand all of the cost components for storage. I’ve seen estimates on tech sites that say that storage costs for a gigabyte’s worth of data have dropped from $9,000 in 1993 to around 3 cents in 2013. Regardless of how accurate those specific numbers are, they demonstrate the huge decrease in storage cost over the last few decades.

But consumers and businesses don’t necessarily see all of these savings, because the industry has gotten smarter and now mostly charges for value-added services rather than the actual storage. Take the backup service Carbonite as an example. Their service will give you unlimited cloud storage for whatever data you have on your computer. Their software then activates each night and backs up whatever changes you made on your computer during the day. This is all done by software and there are no people involved in the process.

Carbonite charges $59.99 per year to back up any one computer. For $99.99 per year you can add in one external hard drive to any one computer. And for $149.99 per year you can back up videos (not included in the other packages) plus they will courier you a copy of your data if you have a crash.

The value of Carbonite is that their software automatically backs you up once a day (and we all know we forget to do that). But that is not a complicated process and there have been external hard drives available for years with the same feature. But Carbonite is selling the peace-of-mind of not losing your data by putting it in the cloud. It must be a very profitable business since the cost of the actual data storage is incredibly cheap. Consider how much extra profit they make when somebody pays them $40 extra to back up an external hard drive.

In the business world, the fees paid for the cloud are all about software and storage cost isn’t an issue other than for someone who wants to store massive amounts of data. One might think the companies in the cloud business are selling offsite storage, but their real revenue comes from selling value-added software that helps you operate your business and manage your data. The storage costs are almost an afterthought these days.

The race to zero is not even close to over. In one of my blogs last week I talked about how using magnetized graphene might increase the storage capacity of devices by a million-fold. That upgrade is still in the labs, but it demonstrates that progress to ever-cheaper storage is far from done. We’ve come a long way from the 720 kb that I used to squeeze onto a floppy diskl!

CenturyLink and the Cloud

Cloud_computing_icon_svgI don’t write many positive articles about the largest US telcos. This is mostly because these are the competitors for most of my ISP clients, but also because the big companies are on the wrong side of issues like net neutrality and privacy. It’s generally pretty easy to find things to dislike about any one of the big carriers.

But I have to say that I am impressed with CenturyLink’s foray into cloud computing. They got into it early and they have carved out a decent market niche. Cloud services is already a huge business and will grow much bigger. I read a recent statistic that says that only about 13% of US corporate data today is stored in the cloud. That leaves a lot of room for industry growth.

The two big giants of the cloud storage industry are Amazon and Google. In fact, Amazon is so large that I read that Amazon has five times more data center capacity today than the next 14 competitors combined, including Google. But I also have read reviews that talk about Amazon as the ‘Walmart of cloud storage’. They are cheap – they have lowered cloud data storage prices 42 times since they started. But they are also somewhat generic and this comes from having a suite of products that tries to satisfy everybody.

But companies like CenturyLink and Peak 10 have created a niche in the cloud computing market by offering customized services. For example, Peak 10 has concentrated on the medical and the gaming industries. CenturyLink cut its teeth on providing services to governments and other large businesses.

There are several components to cloud computing – data storage, transport, computing power and software centralization. Amazon has clearly moved ahead of everybody else in storage capacity, but one has to wonder if this is a long-term advantage. It appears that data storage is moving towards being free, or nearly free. Obviously with the upcoming Internet of Things there is going to be more pressure put onto storage capacity, but the dropping prices for data storage is what has led to the repeated Amazon price cuts

CenturyLink competes much better in the transport arena. They were born out of the merger of Qwest and US West, with Qwest having significant fiber assets throughout the US and the hemisphere. They continued to expand their fiber post-merger and most of the US is close to a CenturyLink fiber. While transport prices have dropped, particularly on the major intercity routes, transport to smaller markets is still a very lucrative business, and having fiber in those markets gives CenturyLink an advantage in many regions.

Amazon also has the edge today in computing power by virtue of owning so many data center assets. Amazon is not ahead only by virtue of sheer number of data center computing assets, but they have also been working feverishly on building faster and more energy efficient servers and switches. This gives them a temporary market advantage, but these kind of advantages usually don’t last too long. There is a major industry shift towards software defined networking and this is going to result in cheap data center routers and switches for everybody.

I wrote this blog as an example that it’s possible for a company to reinvent itself. I don’t think anybody has been thought to be stodgier than CenturyLink for the last decade. While Verizon and AT&T have been adding data customers, CenturyLink struggled with old copper. But CenturyLink is now a player in cloud products and they have recently launched credible new business lines by building fiber-to-the-home networks and also launching their Prism DSL product that is similar to AT&T’s U-verse. I hold CenturyLink out as an example to my clients. If they are able to take the steps needed to make sure that they will be relevant decades from now, then so can any other ISP.

The State of Cloud Services

cloud computing

cloud computing (Photo credit: kei51)

My clients ask me all of the time how they can make money at cloud services. The fact is, for small carriers, there are a few opportunities, but the industry still has a way to go to be ready for prime time for small carriers as a revenue opportunity.

I say this because my average client only has a handful of business customers who can really benefit from using cloud services, and so the small volume they might be able to sell to them does not look like a profitable product line.

What is available today?

First, there is a very robust market in providing data storage and back-up of data. But there is only money to be made in this from business customers because residential customers can get mountains of free web storage if they look around. It’s possible for a residential customer to easily store a terabit or more of data for free.

But businesses don’t want to, and probably should not use cheap or free web storage. There are already horror stories of web storage services that have shut down and that have left people without access to the data they have stored in the cloud. So a business needs to store their data where they know they will always have access to it. This probably means storing it with a vendor that has multiple data centers so that there is a duplicate backup copy of everything to avoid the issue of natural disaster.

And it’s not hard for a small carrier to get into this business themselves and to store some data in their own central office. And if their customers want a second back-up copy there are a number of reputable data centers around the country that are owned by other small carriers and that seem pretty secure and safe. There is even a little money to be made to be the middle man and in sending all of the data to somebody else for your customers.

The other thing that is widely available today as a cloud service is IP Centrex. There are a number of national companies that will sell this service to anybody that has a fast enough data connection.

But one of the catches to this service is that these nationwide sellers do not offer phone numbers everywhere. This means that when they go to sell in rural areas they probably do not have the ability to do number porting to let the customers keep their local numbers. This is a big deal for businesses. We have always assumed that the nationwide sellers work through some other nationwide CLEC to terminate traffic, and those CLECS, like a Level3 have gained the ability to do number portability in RBOC areas, but for the most part they do not have those agreements in place for the rural areas.

But number portability aside, it is possible for anybody to resell the IP Centrex services. If you are competing in a neighboring larger town you could offer these services in the cloud as a reseller of one of the nationwide carriers. The margins are not nearly as good as if you offered this on your own switch, but they are okay.

Finally, the real promise of cloud services is that it could offer the software a business uses everywhere on any device. These are huge advantages to large companies having this ability and many of them have migrated their software to the cloud. But they have done so with a lot of effort. Most companies operate a unique set of programs. While a business may use the standard nationwide software like Microsoft Office or Quickbooks, most companies also run a number of unique and homegrown programs. The real challenge for a company that wants to take its software to the cloud is not getting the big name software to the cloud – because most of that software now has the option of cloud licenses. The issue is in moving all of the home-grown and one-off software that a company uses. As I mentioned yesterday, many companies still operate some PCs with Windows XP. It’s not as automatic to move older legacy systems to the cloud as you would hope and it takes some effort and trial and error to get some things to work in a cloud environment.

And there is no profitable product out there yet for the small carrier who wants to offer cloud software to customers. There are bits and pieces, but no easy platform that just lets you sweep your business customers into the cloud. This is probably coming, but it is not here yet.

So in summary, other than data storage and IP Centrex, there are not a lot of viable, money-making ways for a small provider to make money yet on cloud services. But I think the day when you can is fast approaching. There are bits and pieces already available for offering cloud-based software and the options are growing all of the time.