Amazon (Amazon Web Services, AWS) is the leading enterprise cloud provider. Amazon surprised the market with a larger than expected loss. AWS has had a big impact on the result. What does it mean that the leading cloud company cannot run a profitable business?
Many web businesses are dominated by two or three significant players and we have seen that Amazon, Microsoft and Google are very serious in the cloud business. So it has been a valid assumption that these giants can take the main role in the segment.
Now we have started to see more variables. I wrote in From clouds to fog about how the distributed fog model might come to challenge the clouds. Most probably we won’t see an over night jump to a fog model, but we have already started to see signs that the cloud business will not only be dominated by a few big internet companies.
There are a few factors that seem to become more important when companies are making their cloud selection: 1) security, 2) privacy, 3) global availability and latency, 4) political risk, and 5) hybrid options (i.e. combining private servers and clouds). At the same time we are seeing strong price erosion in the cloud services. AWS has implemented price cuts of between 28% and 51%, and this has now really started to impact their revenue. As one CTO commented, “it doesn’t make sense to agree to a one-year contract with a given discount, when prices go down faster than the natural market pricing.”
Data security and privacy in particular have become a big concern during the last 15 months. It is not only a question of what is the real exposure of data, but also how clients and consumers feel about it. This is also linked to geographical and national borders. Is it acceptable to keep customer data in another country, what if this other country authorities want to get them, what kind of encryption is legal, or are there some countries between our users and the cloud servers that can intercept the data?
Political risk is something companies also consider more. Is it possible that a local government becomes hostile towards us, if we just offer the service over their borders and are not present locally, or do not have a local cloud partner? The political risk doesn’t matter so much for many more casual web services, but if you offer services that are more critical and users expect really high availability, it start to be a relevant consideration.
Latency time is also becoming more important when users don’t expect any delays. This can be particularly relevant in many countries in Asia, if the service is hosted only in the US or Europe, or only in one location in Asia. But it can be relevant even inside China. Local caches are one solution, but they can also raise availability and data security issues.
All these components offer also more opportunities for carriers in this market. For example, Verizon has been active in this market in the US, and it probably has an impact on AWS’ bottom line too. Companies like IBM and Cisco (that also market the fog model) also want to win more market share. AWS hasn’t really lost market, but has also seen it must make heavy investments and offer lower prices. Carriers can have an advantage in this game, if they can build services on their existing infrastructure and customer relationship. And the fast growing use of mobile can allow carriers to offer solutions to have fast access to data everywhere in the network.
Data is in the networks to stay. Designs will probably start to shift from centralized to more distributed architectures. And it looks like cloud players' market shares are not fixed, and neither are the solutions and architectures that will dominate. Clouds are becoming commodities, but at the same time the businesses and technologies have a lot of room for innovations and new business opportunities.