As with other breakthrough developments, startups often lead the way; then the tycoons of the world want to join in on the fun :)
You know you’re in the right field, and that there’s a genuine demand from the market, when the big guys want in!
And what a month it has been for cloud databases: October 2011 saw two announcements from the mega-corporates of the business world- Oracle and Google. Let’s examine what each means for the heating cloud database arena.
The Oracle Cloud
In his closing keynote at Oracle’s OpenWorld Conference in San Francisco at the beginning of this month, the always-controversial Larry Ellison – Oracle’s CEO – announced Oracle’s public cloud.
While this announcement comes a bit too late in the game, it’s an important validation to the proliferation of the cloud: after completely dismissing the cloud a few years ago, the Oracle giant now has to admit it is here to stay.
This announcement doesn’t position Oracle as a market leader, but rather it is seen more as a “me-too” tactic. It’s interesting that even Ellison’s delivery of the news during his talk was phrased along the lines of “now that everyone else has got a cloud, we need one too”…
While there’s still no clear indication when will Oracle’s cloud become available, Ellison describes the new offering to be a mix of PaaS and SaaS capabilities (not that different form Salesforce’s offering), and promises no vendor lock-in (would you believe it ?!). This naturally raised quite a few eyebrows, and flared up its fair share of jokes regarding Oracle’s pricing (“if you pay a million dollars- then, sure, no vendor lock-in” :)). In addition, the Oracle cloud architecture won’t support multi-tenancy (allowing customers to share the hardware but keep their data separate) but rather dedicated machines per user.
Oracle sees its cloud in direct competition with Amazon, Rackspace, Salesforce and other public cloud vendors. But is it really?
If we look at their product – the Oracle commercial database is optimized for single storage (no distribution) and high-end hardware. This simply doesn’t work well in a virtualized environment and with commodity cloud hardware. This, along with its high price point, means cloud users are effectively moving away from Oracle, and doubt they would ever reconsider it as a valid – and cost effective – option in the cloud.
While Oracle is trying to take a (belated) piece of the cloud pie, since it doesn’t re-engineer its database to handle distribution well, basically the Oracle cloud is just another set of physical machines, and the Oracle DB in the cloud isn’t really a cloud database- but simply the same old RAC installation- suffering from the same old problems.
The reason Ellison is so against multi-tenancy is because Oracle doesn’t support it and it’ll take them a quite a while to develop it. Using multi-tenancy as “big-bad-wolf” made-up argument to supposedly strike fear in perspective clients to get them to pay more for the dedicated option won’t fly. Done right, true multi-tenancy and properly designing the database for a virtualized framework to begin with – actually adds value, rather than risk. Many times, sharing the hardware is your best bet for getting your infrastructure costs down; scale-out cost effectively and enjoy unlimited throughput and computing power when you need to without breaking the bank.
What will happen to Oracle’s cloud only time will tell, but as we stated before: in the cloud, big names don’t necessarily mean big value.
Google Cloud SQL
So Google is also seeing potential in a relational database designed for the cloud.
Makes sense :)
This is very good news for cloud databases, as yet another indication that despite the NoSQL hype, SQL is not going anywhere. While NoSQL may be most suitable for heavy numbers-crunching and BI applications, SQL is still the preferred choice for 99% of use cases – so there’s better be a way to ensure it runs smoothly in the cloud.
Google’s Cloud SQL is still in beta, but the offering seems like a shared environment that’s a combination of Salesforce’s Database.com and Amazon RDS – suffering from both platform’s limitations. The key benefit to Google’s SQL offering is its built-in integration with Google’s App Engine, but here also lies its limitation – as it’s limited to those running on Google’s cloud.
Another key limitation, and possibly the most important one is that MySQL scalability in the cloud is a major issue still not addressed.
Where Google’s and Oracle’s new offerings would end up and how they would stack up against other database capabilities for public clouds remains to be seen. The more products and unique capabilities we have out there, the more choice users have to find a good fit for their specific use case. One thing is certain: it’s going to be interesting playing with the big guys!