451 Research: Cloud Price Index: mapping the ‘cloud war’ battle lines

InsightaaS: A cloud service is a service, is a service. Right? Not so, according to the 451 Research Group, which has developed a ‘Cloud Price Index’ to gauge cloud costs and identify opportunities for provider differentiation. In the report below, 451 senior analyst, digital economics, Owen Rogers offers a recent sample from the Index – “like consumer price indexes… a basket of goods [made up of] the services required to operate a typical Web server application” – that considers price gaps between services provided by “hyperscalers” AWS, Google and Microsoft. Rogers argues that commoditization can be found in services for which price gaps are narrow, as well as the obverse: wide gaps in pricing signal areas that are ripe for differentiation, a sound thesis that delivers few surprises, at least on the commodity side of the equation.

So where are the best opportunities for differentiation as per 451 research data? “Up the stack” and beyond the provision of virtual machines and object storage is the answer, though conclusions drawn in the report about wider price gaps and service differentiation are more problematic than the identification of commoditized services. For these services, Rogers rightly asks additional questions of the data results – are block storage and load balancing really areas of differentiation, or simply not on customers’ radar? If it’s true that “a GB/sec of bandwidth is a GB/sec of bandwidth – adding more value is difficult,” why is there a large gap for the bandwidth variable? Is the largest gap – in nonrelational NoSQL databases – really a function simply of “novelty factor”? Ambiguity on the service differentiation side may be in part a function of applying a metric (pricing) designed to measure commoditization to complex service categories that defy strict definition, and in part a product of a service’s relative positioning on the Index scale. Herein lies the best value in the Index: if some questions linger in review of a price gaps snapshot, trend analysis in the Index service over time should better reflect changes in a highly dynamic market and the drift to or away from commoditization for different service categories – information that may help educate cloud service buyers looking for value, and providers looking to refine services offerings.

 

Owen Rogers, senior analyst, digital economics, 451 Research Group
Owen Rogers, senior analyst, digital economics, 451 Research Group

AWS has made significant price cuts to its cloud services over the past few years, forcing the likes of Google and Microsoft to fight back by promising to match these cuts, too. Services that the cloud providers are forced to keep pace with price-wise are slowly commoditizing as price, not product differentiation, becomes an increasingly important decision criterion. But cloud is more than just virtual machines – there are other services where cloud providers might see opportunities to differentiate in other ways, thereby deriving greater margins. On which services are the hyperscalers matching their competitors, and on which do they see greater differentiation than price alone?

The 451 Take

A so-called price war between cloud providers is not being fought on every front. Data from the Cloud Price Index shows that virtual machines and object storage may be services that cloud providers feel the market price must be keep pace with. But cloud is more than just a single service; the term cloud represents a varied array of services, with each having its own pricing strategies and tactics within the walls of service providers. Since some cloud services vary in price substantially between providers, we are not seeing the whole industry commoditize. The similarity in price between virtual machines and storage suggests that these are commoditizing, but then again, the variation in others suggests there are still opportunities to gain greater margins and to demonstrate value to the customer. Gross margins on some services are likely to decline, but because of the opportunity to upsell other services, this does not mean service-provider net profit margins are likely to decline also. For service providers, the opportunities for greater margins now lie further up the stack; enterprises should ensure that the value they are getting for paying more for simple IaaS is clearly identifiable.

There's been much debate recently as to whether cloud is a commodity. Providers are keen to avoid this word, associated with smaller margins and a generic, cloned user experience regardless of provider. Of course, providers want to show their services offer value – the ability to transform or streamline a process, department or even an organization in a unique manner allows a higher price to be charged, meaning a greater gross margin. But there are signs that some cloud services are commoditized: Price cuts by the likes of Google Compute Engine, Azure and AWS show that price is an important factor in consumers' decisions. And how different really is one virtual machine from another?

To evaluate a market as broad as cloud based purely on the prices of virtual machines and storage alone is foolhardy. The typical cloud provider doesn't just offer these basic services – it offers a huge range of services, including database, CDN and SaaS. As shown in our Cloud Wars series, each of these services has its own differentiated features, value proposition and pricing. And perhaps more importantly, the combination of services and their integration varies from provider to provider. So although virtual machines are perhaps becoming more price-sensitive and less differentiated, that is not necessarily the case for all cloud services and cloud providers. If a service is exactly the same from provider to provider, it is essentially a commodity; a purchaser can't make a decision on which to choose without relying heavily on price. If a service varies in its features from provider to provider, it is differentiated; buyers will pay for the features they need.

To examine the 'real world' cost of cloud over time, 451 Research has created a Cloud Price Index. Like a consumer price index, our Cloud Price Index is made up of a basket of goods, but in our case it is a specification of the services required to operate a typical Web server application. Changes in the CPI over time will reflect how changes and differences in pricing in the cloud industry are reflected in real-world use. The Cloud Price Index is the specification of a multi-service, three-tier cloud application consisting of Linux virtual machines, object storage, block storage, relational database, NoSQL database, load balancing, access control lists and snapshot backup in a resilient architecture.

So, among the hyperscalers – AWS, Google and Azure – which services are similarly priced across the board, and which are far apart? By examining this issue, we may be able to identify where cloud providers are keeping pace with the rest of the market, and where they believe they have something worth paying a little more for.

Mapping the battle lines

The following chart shows the range as a percentage of the average, which we are using as a measure of pricing similarity between providers. Using data provided voluntarily by AWS and Google as part of the October edition of the Cloud Pricing Index, and an estimate for Microsoft Azure, we are able to compare the costs of delivering the same services from each provider. A smaller value indicates similar pricing; a larger value suggests a wider discrepancy.

451 cloud price Index

 It is no surprise that the hyperscalers are keeping pace with each other for object storage and compute. These battlegrounds are the most aggressively fought over, with cloud providers slashing these prices on a biannual basis and the media being keen to cover the announcements. They are also fairly undifferentiated between providers: A GB of storage is essentially the same, regardless of object-storage platform, and similarly sized virtual machines with a Linux OS installed on them are likely to perform similarly. From the service provider's point of view, it also makes sense to keep these services inexpensive – storage and compute are likely to be the first services enterprises dip their toes into, and can provide volume from which to upsell other services with greater gross margins. Relational databases are also similarly priced between cloud providers, as are DNS services. A relational database service is essentially a virtual machine with a preinstalled database; it's difficult to add value to something so simple, so it's no surprise providers need to keep pace with the market price. DNS services have been around for a long time, so there is limited differentiation here also.

Block storage is a bit of a surprise. One might think block storage, which is essentially nothing more than a hard drive, might be fairly standard. Perhaps there are discrepancies in the performance that providers think might make their offering stand out. Or perhaps block storage isn't even on consumers' radars when it comes to cost. The same issue applies to load-balancers – is this really an area of differentiation? It's more likely this isn't a primary consideration to consumers when building applications.

Cloud providers argue about network and capability, and port connections, contention and LAN bandwidth vary from provider to provider. Some state that having one's own network guarantees a certain level of performance. But at the end of the day, a GB/sec of bandwidth is a GB/sec of bandwidth – adding more value is difficult. But in our specification, cost varies largely between providers. The broader service providers – particularly carriers such as Verizon, AT&T, CenturyLink and Vodafone – emphasize their network ownership as a key differentiator.

Support also varies by substantial amounts, with the different providers showing value in support as standard and for a chargeable amount. This isn't a surprise – support is the most basic form of value-add there is. As well as support services varying in terms of response times, credentials, expertise, reputation and location, they may also be bundled in with a service or charged separately, acting as portfolio and organization differentiators. Support services upgrade IaaS to be closer to an outsource arrangement, giving service providers better margins, stickiness and a relationship on which to build further sales.

But the biggest price range variation in our basket of goods is in nonrelational NoSQL databases; the variation in NoSQL database pricing is 10 times the variation of object storage using our measure. Is this because NoSQL is a newer technology that enterprises are only starting to use, or does each cloud provider think its service is unique in the market? We think it may be because of the maturity of this technology. Compute, storage, bandwidth and RDS are all typical elements of an application, even before the cloud came along – enterprises are not only comfortable using these, they are happy to experiment with different providers. NoSQL, on the other hand, is a newer technology, which enterprises might be willing to pay more for since they have no comparable experience in using it; perhaps 'novelty factor' sums up the premium payable. Enterprises are only just dipping their toe into NoSQL – they don't really want to have to think about comparing providers right now.

Conclusions

Of course, this is only a sample of three, and there are slight discrepancies between competing services. Furthermore, we are looking at a fixed specification – in other applications, the variation might change substantially. But this does provide us with a guide. The most important takeaway is that any so-called price war between cloud providers is not being fought on every front. Yes, cloud providers may feel they must keep pace with the market price of virtual machines and object storage, but cloud is more than just a single service.

Some services, like virtual machines and storage, perhaps, are commoditizing, but then again, there are other services that the cloud providers see that are adding greater value. Gross margins on some services are likely to decline, but because of the opportunity to upsell other services, this does not mean service-provider net profit margins are likely to decline also.

The lesson for service providers: The opportunity no longer lies in simple IaaS – one must look further up the stack to differentiate. The lesson for enterprises: If you are paying more than the market price for simple IaaS, make sure you can justify what additional value you are getting.

 

 

LEAVE A REPLY