Multi-Tenancy Debate Won’t Go Away
Under more normal conditions I react to blogs on a low key basis. If I like something I might not even mention it and the same is true if I dislike or disagree with one. The blogosphere is a big place and it is unproductive to be running around all day saying “Yes, I like this” or “No, not so much.” But I am grateful that so many of you read and even comment on what I post.
However, when a vendor weighs in on an official site and clearly makes a statement that I think is wrong and or self-serving, I feel a need to provide a counterbalance. Earlier this month Meg Bear, Vice President, Oracle Cloud Social Platform posted, “Multi-Tenancy and Other Useless Discussions” and from the headline I think you know where this is going.
The whole multi-tenancy debate of the last decade has been run by the same logic that a five-year-old uses at dinner — don’t let the peas touch the mashers or the chicken! Ironically, the chocolate syrup does indeed touch the vanilla ice cream and that is socially acceptable, but I digress.
According to Bear, the debate is no longer relevant not because we have all grown up and understand that, just as a bank co-mingles everyone’s money keeping yours separate from mine by use of metadata (account number, balance, statement, things like this), our data can be co-mingled without worry about cooties. No! Bear’s rationale is that Oracle has a superior “modern application architecture.”
Can we please be serious? I beg to differ for the following reasons.
IT is over. Note I didn’t say finished or in any way imply irrelevant. IT is over as a disruptive innovation and driver of the economy. Some macroeconomists who study the long economic wave a.k.a the Kondratiev Cycle have begun noting that the IT wave is over — just look at the economy today and you know this. IT is the economy in much the same way that cars and petrochemicals, once big drivers in their own right, are the economy. They don’t grow a zillion percent any more and now neither does IT. Over, finito, gonzo. Bring on the next Kondratiev Cycle, please, and hurry up!
But quickly back to Ms. Bear. IT is rapidly commoditizing — that’s the real message of cloud computing. The cloud frees up budget once spent on hardware for more productive pursuits. Under those conditions, dedicating spindles that remain half empty for everybody’s peas, mashed potatoes and chicken and dedicating separate database instances is both wasteful and unprofitable for the vendor. It also doesn’t seem very modern to me.
Fortunately, Ms. Bear assures us equally that the Oracle Fusion HCM applications are multi-tenant and that it doesn’t matter. If it doesn’t matter then, “the lady doth protest too much, methinks,” thanks to Mr. S.
So what is the point? Apparently it’s business value as in the value the applications enable the business to realize through their use. But according to Bear, multi-tenancy is not an enabler by, say, enabling greater throughput, faster processing and quicker searches over fewer spindles; things like that, because multi-tenancy doesn’t matter.
And that’s because…?
Discussion
Speak Up
You must be logged in to post a comment.
Almost by definition, every SaaS is multi-tenant. At least for all of us that create software for businesses in the CRM space. We all have databases, and they are (or they should be!) kept separate, and it wouldn’t make sense for us to have a hardware dedicated to each individual client. It’s the reason why Amazon’s world of virtual servers is so popular. But of course, sometimes a spade isn’t always a spade (or to use your example, sometimes people like caramel on their ice cream).
Security is important, and in Amazon, where there have been many documented cases of servers getting rooted and other virtual servers getting hacked, this is not a trivial matter. Nor is it if you have your own servers – just ask Epsilon, Silverpop, Campaign Monitor, or any number of highly publicized hacked SaaS providers.
In so many ways, IT is becoming a commodity, but I wouldn’t go so far to say it’s completely over. I think there are a lot of parallels we can draw with other mature industries – automobiles, phones, bicycles – all have been around for a long time but continue to produce amazing creative changes. Innovation is still the lifeblood of many SaaS companies, and will make the difference between the larger, more stagnant companies, and smaller more agile service offerings.
In my very humble opinion, after living through the dot-com boom, bust, and whatever you call this next phase, multi-tenant architecture is in many ways just another name for what we’ve been doing for a long time. Just like “cloud computing” suddenly became a new name for what we’ve been doing for the past 15 years and everyone wants to be “on a cloud” now. Not every multi-tenant system is built the same, and not every cloud is built the same. But innovation can and will still take place, both architecturally and functionally. It’s still relevant, just like we want to make sure our car starts every day, but as technology progresses it becomes more something we take for granted and assume the SaaS providers have all figured out.
That’s precisely the point. Older economic waves don’t vanish, they become assimilated into the economy and assume characteristics of the broader economy. Thus, cars and petrochemicals, just to pick a couple, are still vitally important but neither has the annual doubling that an industry or sector in rapid growth does. It’s the rapid growth that signals the disruptive innovation and without it, or when it slows, the innovation can be said to no longer be in ascent and thus, it is “over” not as an important idea but as a disruption. The difference is important since the rapid growth sector is what drives company formation and job growth. Look at cars and petrochemicals and you see this point. High tech and IT have now reached this point too. Our devices are hundreds of dollars (cheap even by historical standards) and they are made off shore. They are not providing jobs that people in first world economies want because they don’t pay enough to generate a middle class income. Our software is cheap and near free. Many of the applications that run on my “designed in California, assembled in China,” $500 tablet are free or cost 99 cents to $4.99. The “real” software we use in business costs tens of dollars per user per month. Importantly social media is free and few have figured out a business model for it that makes money. Again not the kind of money that drives innovation or middle class lifestyles. Compare this to the 1950s when a megabyte of RAM cost a mega-buck. IT is over. It’s long passed the time to think about what’s next.