Forrester Confuses “Best” with “Most Accessible”
The Cloud Zone is brought to you in partnership with Iron.io. Discover how Microservices have transformed the way developers are building and deploying applications in the era of modern cloud infrastructure.
I’ve got a love/hate relationship with the traditional analyst firms. Their people are incredibly smart and very thorough, but sometimes in their search for massive levels of details, they miss the very point of technology. A good case in point is the recently released Forrester Enterprise Cloud Database Wave report. For those not up on the Forrester Wave reports, for those who haven’t come across them before, is another version of Gartner’s Magic Quadrant methodology – essentially it classes products based on it’s current offering, and on the stated product strategy to differentiate products within a category.
The Cloud Database Wave findings can be seen below:
So far so good right? Except when you dive into the methodology. Let’s take a look at one of the categories upon which Forrester scored products, that of multitenancy. Multitenancy should be a very simple test – either a product IS multitenant, or it isn’t. There wouldn’t seem to be any gray areas in there right? Well, actually no. The Forrester scoring goes like this:
Describe the product’s multitenancy model to support cloud databases. Is the cloud DB offering a multitenant implementation, or is the offering a dedicated database running in a VM? Does the vendor support both modes? Describe how customer data can be isolated (i.e., into specific servers, storage, partitions, shards, schema, table or rows). Can customers put their data on dedicated infrastructure on the public cloud? Does the service provide “dedicated compute/memory capacity”? Are customers able to view metrics such as disk I/O and read/write latency?
Forrester gave products a multitenancy rating as follows, with 5 being a perfect score:
5 = Natively supports both dedicated full database platform running in a VM environment and multitenant database. Supports complete isolation of customer data into specific servers, storage, partitions, shards, schemas, tables, and rows in a multitenant database. It also allows customers to put their data on DEDICATED hardware that’s separate from other customers on the public cloud. It also allows for dedicated compute and memory capacity and resources that are separate from other customers. Customers can view metrics such as disk IO and read-write latency using tools.
3 = Supports either the dedicated full database platform running in a VM environment and multitenant database. Supports complete isolation of customer data into specific servers, storage, partitions, shards, schemas, tables, and rows in a multitenant database. It also allows customers to put their data on dedicated infrastructure that’s separate from other customers on the public cloud.
1 = Supports either the dedicated full database platform running in a VM environment and multitenant database. Supports complete isolation of customer data in a multitenant database.
0 = Supports no multitenant database facility.
So essentially Forrester is determining multitenancy is a worthy category on which to score products, and then also determining that products that a completely focused on multitenancy are sub optimal. I mean I kind of get what they’re saying here – the products that offer the widest selection of deployment options to customers are preferable, but in scoring down multitenant-only products in this classification, they are making a judgment call by saying widest applicability trumps most focused multitenant execution. This perspective is even more strange given one of the key takeaways they came up with:
Cloud databases are Not a database in a Virtual Machine – Cloud databases are fully automated multitenant services that present a database capability but are managed under the covers, giving, to varying degrees, elastic scale, performance, and availability management and programmability on a pay per use basis
So the key takeaway that Forrester came to, namely that a cloud database shouldn’t be thought of as a database in a VM, was completely disregarded in giving pureplay multitenant vendors a lower score than those who deliver database-in-a-VM functionality. Does. Not. Compute.
So what’ll it be Forrester, will you back yourselves when you say that a real cloud database is one which is fully automated and multi tenant, or will you continue to be an apologist for the all-things-to-all-people crowd?