Blog Post

Why Cloud Computing Sells and NoSQL Is Fading

Three years ago, I spent a few post-Thanksgiving hours thinking about how vendors were defining their relatively nascent cloud computing efforts. Specifically, I was intrigued by IBM’s (s ibm) Blue Cloud offerings (remember that term?) and the notion that the first product under that banner would be an on-premise, blade-based appliance (which ultimately materialized in the form of IBM CloudBurst). Cloudwashing had yet to reach critical mass, and I wondered whether applying the cloud label to everything was a wise idea. In the meantime, marketers correctly decided it was, but concern over the wisdom of overusing buzzwords hasn’t gone away.

Cloud Is a Winner

To be fair, cloudwashing has proven itself an effective strategy. Even former cloud computing hardliners like myself have come around to the notion that anything delivered via the Internet, promising on-demand resources or simplifying provisioning via self-service can fairly call itself cloud. SaaS, virtualization management, some forms of managed hosting, magazine printing — if you want to call it cloud, have at it. I won’t argue. Nor will investors: In 2009, startups using cloud or SaaS in their business descriptions raised $374.77 million in venture capital, almost twice what they raised in 2007. Since September, Cloudera, RightScale and Joyent raised $65 million by themselves, and during a two-week stretch in November, four cloud-storage startups — Zetta, Nirvanix, Gluster and Scale Computing — combined to close $47 million.

The reason cloudwashing is so effective, I think, is that the cloud works well as an icebreaker. Not only is it a buzzword, but the potential savings in cost and improvements in flexibility are well-known and very real. Calling something a cloud solution will get your foot in the door, then it’s up to the vendor to get into the details of how it actually works. Even if a product doesn’t perfectly fit any given individual or corporate definition of cloud computing, results are, ultimately, all that matters. Call it whatever you want, as long as it works.

Actually, there’s a real risk involved with not invoking the cloud if the term might accurately apply. Particularly for advanced IT products, leading with descriptions of what they actually do can just end up confusing customers or failing to even attract their attention. I’m thinking of companies like Appistry, which experimented with terms like “hive computing” and “application fabric” before settling into its cloud groove, or, which wasn’t getting nearly the attention it deserves under its previous VMOps moniker. You’ll also notice how quickly VMware (s vmw) dropped reference to its Virtual Data Center Operating System.

Cassatt was ahead of its time in selling cloud-like capabilities, but because cloud hadn’t yet caught fire as a marketing tool, the company was left trying to sell its Cross-Virtualization Manager. Companies didn’t really get why they should care about dynamic resource allocation, full-scale automation and intelligent power management. By the time Cassatt began marketing itself as a cloud computing vendor, it was too late. Faced with bankruptcy, the company sold its assets to CA Technologies (s ca) in May 2009, and is now an integral part of CA’s cloud portfolio.

Is NoSQL the Next Grid Computing?

Cloud computing might be one of those terms that’s so transcendent and so widely understood (even if only in theory) that there’s little risk in using it. Three years in, cloud interest is only climbing. On the other hand, one of cloud computing’s predecessors, grid computing, got pigeonholed early as an IT model only applicable to high-performance computing, and vendors trying to sell more widely applicable grid products were forced to jump marketing ships in a hurry. Some, such as DataSynapse, tried out tongue-twisting terms such as “application virtualization,” “real-time infrastructure” and “dynamic application service management.” When Tibco (s tibx) bought DataSynapse, it had another term for those products: cloud computing.

The same thing might be happening to NoSQL. CouchDB commercializer CouchOne recently dropped NoSQL from its marketing vocabulary in order to position itself as an ideal data and development platform for mobile applications. According to CouchOne, NoSQL has become so associated with big data that it’s difficult to convince customers that non-relational databases can be used for more-mundane tasks. It’s hard to argue CouchOne’s decision. MarkLogic, technically a NoSQL database, has never used the term and keeps far away from the NoSQL crowd, and it has built a solid business as a content-management database and unstructured data store across the media, government, financial services and general enterprise industries. It’s still really early, but two vendors actively avoiding references to NoSQL might be a sign of things to come, and a sign that not every potential customer thinks their situations meet the definition of  Big Data.

Not that there isn’t risk involved with bailing from the NoSQL bandwagon. If NoSQL isn’t the next grid computing, CouchOne might have lost a large number of customers for which the term was just a jumping-off point, common ground on which to start a discussion (although it did tie its horse to another hot area — mobile). If it’s doomed to forever invoke big data, though, other NoSQL vendors had better start considering their options. Maybe they’re fine riding the still-cresting big data wave, or maybe they have other ways to sell their products (hint: leading with distributed key-value database isn’t the best way to spark a conversation outside Silicon Valley). If all else fails, there’s always cloud.

Related content from GigaOM Pro (sub req’d):

3 Responses to “Why Cloud Computing Sells and NoSQL Is Fading”