In a conversation this past week the question was asked are cloud providers the CLECs of the twenty-tens? It is an interesting question to ponder especially if you accept that the “cloud” is the basis for 453,000,546,001 investment ideas on Wall Street. Moments before the markets opened on Friday December 9, Jim Cramer and Melissa Lee must have said the word cloud ten times on CNBC. I am clearly on the record as a bit of a cloud bear.
In May 2011 I wrote “Every few years the tech industry manufactures a new buzzword and then works itself up into lather. If it sounds really big, then Wall Street notices and bunch of people create baskets of themed stocks. Cloud is the new winner. We have cloud stocks, cloud baskets, cheap ways to play the cloud, blah, blah, blah. All this cloud talk is really about two elements. Content distribution and the compute point. I am using content as a catch all term to include digital media and apps. Where is the content stored and where is the compute point in the network. That is what we are really talking about when we are talking cloud and when I was looking at network designs in late 1980s and early 1990s we were doing the same so I must been doing cloud networking 20+ years ago. We had software as a service in the 1980s it was called SNA. I competed with service providers who were outsourcing networks in the 1990s so that must have been infrastructure as a service. All we are really doing is putting content and compute back in the data center after spending 25 years distributing it. I am just wondering if we distribute it again, but in the form of smaller DCs to augment the massive warehouse DCs. That is the thesis I am thinking about at this point.” I reviewed my June thoughts on the cloud and I see no reason to make any changes. Six months later do I think cloud providers could be the CLECs of twenty-tens? I think it is possible because I find it hard to frame these data points that are often orthogonal to each other.
Let us start with the mega trend hypothesis. The belief here is that computing (I am using the term computing in this blog post as a catch all for compute, storage, networking) is going the way of the public utility. In the long run, it will be cheaper to deploy computing on a mass scale and this will support the rise of a several large cloud (i.e. computing) providers who will deploy ware-house scale data centers to support the needs of individuals through large corporations. As with your local utility, it is cheaper and easier to contract for energy than to build your own power generating sources. The mega-scale cloud provider will drive the cost of computing to the point in which is wasteful to own in-house.
I have spoken to a number of enterprises in the F1000 range as well as to colleagues who sell to this market and I have yet to find anyone who is abandoning their private cloud in favor of the public cloud. I am sure there is some company out there; they are just really hard to find so I wish they would standup and shout. I have seen enterprises testing new apps in the public cloud, but production level apps are moved to back into the private network and distributed via the private cloud. I think the reasons for this are many ranging from: trust, security, compliance, control, ownership and it feeds the reptilian. Intellectually I understand the cost models, the total cost of ownership arguments, but I think corporations feed their egos. That is why they buy Superbowl ads and that is why they all wanted to buy mainframes in the early days because if your company could afford mainframe, it put you in a different class. I think the same will be for the private data center. Corporations will still want to have them because they can; it kind of feels like the executive jet market.
I can clearly see the small medium business (SMB) market embracing the public cloud. These are organizations with little IT infrastructure and probably cannot wait much longer to dump what they have for a cheaper and more sophisticated IT platform from a cloud provider. This was supposed to be the market that would sustain the CLECs of the 90s and that did not end well. The question that should be considered is how many public cloud providers can be sustained by this market segment and do they need to be part a full service provider?
Another point to consider is on what basis will cloud providers compete? I think the answer to that question is price and price is not really a good competitive dynamic for the industry as a whole. Will public cloud providers become regulated? Will states or municipalities create cloud provider utility commission boards to set rates and levy taxes? If game is to get to Wal-Mart scale for cloud computing, then we are clearly assuming Wal-Mart margins. That will be a problem for many people, but probably a boon to the average consumer. I inserted Perez’s model on technology market evolution to left from a prior post. I would state that the institutionalization of IT or the evolution to utility computing would be supported by her research.
When we look back on the CLECs of the 90s, we must remember what happened. I spent a year or so writing a book on this subject and I will now attempt to condense it to several sentences. Post the Telecom Act of 1996, CLECs were formed to offer lower cost local services. Unfortunately, they had to provide these services over the existing infrastructure that was owned by the RBOC/ILEC. Long haul, wholesale and dark fiber providers were relying on CLECs to be an important source of local traffic (consumer and enterprise class) as the market deregulated. As the market evolved CLECs failed because competing on price was difficult and the technology advancements did not equate to significant profitability and a compelling value proposition over incumbents. RBOCs, ILECs and MSOs owned the local loops or last mile connection and the high end enterprise market stayed loyal to incumbent service providers. Eventually the long haul, dark fiber and wholesale providers either (1) merged to avoid a Pyrrhic victory in pricing or (2) acquired CLECs to capture local loop revenues. Fifteen years on from the Telecom Act of 1996 we have a telecom services market that looks more like pre-1995 or the pre-1984 breakup, than the market the framers of the Telecom Act of 1996 had envisioned. Does this mean we need a Cloud Act of 2013? That is part of the net neutrality debate and best addressed in a separate post in the future.
My observations of the cloud/data center market tell me that new companies funded or created in the last few years clearly choose to put a lot of IT infrastructure in the cloud. The public cloud is cheaper, faster, easier, does not burden the company with unnecessary IT assets and typically the corporate culture in young companies is risk on, not risk off.
I think an important question to consider is how important is content ownership, content security and content control? I think the answer is different across the spectrum from the individual to the F5K enterprise. If you are an individual consumer most likely you have chosen to use Google Cloud, or Apple iCloud or some other cloud service for your personal data. I would observe that hosting your content in the public cloud means you are comfortable or perhaps unaware of governmental requests to see your data and feel comfortable with security around potential hackers trying to steal your data. I think answer to how comfortable we are with content in the cloud is varied. I think individuals are the comfortable using the public cloud (e.g. Gmail, iCloud,Amazon Cloud Drive). I the SMB market is a little less comfortable, but open to public cloud services (e.g. Amazon S3, Google Cloud computing, Savvis Cloud and other cloud services) compared to the individual user. The market that I think is not ready for the public cloud is the larger enterprise market. They will use public cloud technologies and techniques, but deploy them internally in private clouds.
Summarizing my view of the state of the public cloud is that individual users are most adopting of the public cloud and that usually comes as part of broadband service or wireless service or application service (i.e. digital lockers). For this market, mega data centers have the scale to win. For the SNB market, mega data centers have a play, but I think small to medium containerized or distributed DCs will be the winner. For the large enterprise, I think they will continue to own their own DCs and eschew the public cloud.
The question to consider in terms of public cloud adoption is how critical is controlling or owning your compute, storage and network elements? I think the answer varies. At the individual user level it is not that important, but as the size of the entity increases the answer changes significantly and the larger the spending entity the more uncomfortable they are with transitioning assets to the public cloud.
Putting it all together, I think the market is evolving along a line that does not support a wave of massive outsourcing to the public cloud for ten to fifteen years. I think the Fortune 3K and maybe as large as the F5K see value in owning private data centers. I am on record as saying that I have seen a trend in which corporations see value in running their own network, which means contracting for dark fiber and lighting it themselves. In this evolution of the network, large enterprises look like a private service provider and what they desire is peering relationships with public service providers and other companies (i.e. business partners) acting as private service providers too. If that assumption is true, then the private data center will not go away for a long time. Of course I could wrong and we could someday all be using the great public computing utility in the sky.
I do think that content does matter and the data center will be important and this may lead to cloud providers looking more like fully integrated service providers, but that is reserved for my next post in a few days. I am off to CA for a few days and hope to have the next post which is an extension of this post up by the end of the week.
* It is all about the network stupid, because it is all about compute. *
** Comments are always welcome in the comments section or in private. **