by Bernard Golden

The State of Cloud Computing in Japan

Nov 05, 2009
Cloud ComputingVirtualization

What can you learn from a conversation with Japan's second-largest telco, which is also a cloud service provider? Plenty, says's Bernard Golden.

During this week’s Cloud Computing Expo, I had an opportunity to meet with a new Japanese Cloud Service Provider. This company, KDDI, is the second largest telco in Japan, and offers a wide range of network-based services, including colo and managed hosting. It has now moved into offering IaaS cloud services (starting roughly six months ago), basing its offering on well-known cloud infrastructure software company 3tera. I found our conversation really interesting, for the following reasons:

KDDI chose 3tera because of its “superior GUI,” which sounds like a bit of an odd reason. In following up on the discussion with 3tera, it seems that superior GUI was shorthand for 3tera’s GUI driven dynamic, scalable provisioning. In other words, 3tera provides the ability for end users to directly provision their own instances, with 3tera software dynamically scaling the application infrastructure as needed.

If you look at my company’s cloud services taxonomy diagram, you can see that we place agile provisioning and scalability as a core user-level service. Too many IT organizations define cloud computing as improved operations, rather than treating it for what it is: a way for end users (i.e., the application groups) to control their own infrastructure. 3tera’s product enables end user provisioning, which KDDI clearly recognized as necessary for its offering.

[For timely cloud computing news and expert analysis, see’s Cloud Computing Drilldown section. ]

Another really interesting item was the typical company profile KDDI identified as currently using the cloud offering. According to KDDI’s experience, enterprises are not really using the cloud to create dynamically scalable apps, but instead treating the infrastructure as dynamically provisioned hosting. It specifically identified dynamic provisioning of clustered RAC as a use case for enterprises. In other words, enterprises are not yet using the cloud characteristics of the KDDI cloud.

However, system integrators (SI) are using the KDDI cloud as an infrastructure for “cloudy” apps. KDDI described how SIs are using the KDDI cloud to create SaaS applications, although it wasn’t clear to me if that means that SIs are porting packaged software products to SaaS capability, or if SIs are moving enterprise apps into the cloud and calling them SaaS. This ambiguity might be ascribed to the fact that the conversation was carried on through a translator, who, though she seemed pretty tech-savvy, might have been somewhat inaccurate when translating the new concepts of cloud computing from one language to another.

When queried as to whether enterprises were making any use of the KDDI cloud for “real” cloud applications, the gentleman from KDDI mentioned experiments and pilot implementations. It seems that, despite the cloud offering coming from a well-known provider like KDDI, enterprises are moving very deliberately toward embracing the cloud. One thing I neglected to ask was the actual uptake of the KDDI cloud offering; it always seems like the really interesting questions present themselves after the conversation!

One thing the KDDI representative said was absolutely consistent with what I’ve heard from many people here in the US. I asked him about Amazon’s cloud presence in Japan and he characterized the primary users of Amazon’s cloud as consumers. If you examine the chart I presented in a recent blog posting about the number of EC2 instances being instantiated each day (around 50K), it’s clear that there is a ton of use, which is unlikely to be completely consumer driven. Frankly, many tech market participants find it difficult to believe that “serious” computing is being done on AWS and therefore tend to dismiss it as a “consumer” offering. It seems that those that dismiss Amazon in this fashion find it undignified that a non-tech company could enter a market that the traditional players regard as their own. I am absolutely confident that if one were to examine the Japanese-based users of AWS, one would find many significant organizations consuming AWS resources.

The problem is exacerbated by the fact that use of AWS is relatively anonymous; Amazon itself really has no idea who is using its IaaS service. In this respect, AWS resembles open source software. For a long time, many software vendors dismissed the extent of open source use within enterprises, overlooking the fact that the anonymous distribution characteristics of open source meant that no notice had to be given by a using organization; that is, the open source user had no need to inform a vendor that it was using the software. Consequently, the user base of open source software tended to be anonymous. This led to the understandable, but myopic, conclusion that therefore no big organizations were using open source software. The subsequent developments illustrated that anonymity does not equal non-use; we’re seeing the same phenomenon with Amazon’s cloud, with, one can predict, similar outcomes: transformation of the user base to the disfavor of the incumbent providers.

Going forward, KDDI is likely to be surprised by how quickly enterprises begin using its service to create “real” cloud applications, well beyond the previously described dynamic hosting provisioning. It will be interesting to see how the KDDI cloud develops—especially because one of the reasons organizations embrace cloud computing is due to dissatisfaction with the economics of colo or managed hosting.

To the extent that customers embrace the KDDI cloud, this might disadvantage its mainstream hosting business. Certainly we’re seeing mainstream organizations beginning to undertake cloud computing initiatives with a real seriousness of purpose. The ongoing financial crisis and consequent pressure on IT organizations to cut costs has caused senior IT management to recognize that the future can’t be business as usual with a few trims here and there. A new approach that ties costs much more closely to outcomes is required, and it’s just a matter of time before the same dynamic plays out in the Japanese market.

Bernard Golden is CEO of consulting firm HyperStratus, which specializes in virtualization, cloud computing and related issues. He is also the author of “Virtualization for Dummies,” the best-selling book on virtualization to date.

Follow Bernard Golden on Twitter @bernardgolden. Follow everything from on Twitter @CIOonline