Multi-tenancy vs. single tenancy: Looking beyond TCO

Multi-tenancy, single tenancy...the technical debate rages. But are we concentrating too much on the TCO argument and not seeing clearly into the future?
Written by Dennis Howlett, Contributor on

Some of my colleagues kicked off a furious discussion. Josh Greenbaum, Phil Wainewright and Bob Warfield are staking out the multi-tenant v single-tenant battlefield. I'll lob in my two penn'orth.

I was sitting next to Josh at the Workday event as he fretted over whether multi-tenant brings user versus vendor benefits:

While the benefits that multi-tenancy can provide are manifold for the vendor, these rationales don’t hold water on the user side.

That is not to say that customers can’t benefit from multi-tenancy. They can, but the effects of multi-tenancy for users are side-benefits, subordinate to the vendors’ benefits. This means, IMO, that a customer that looks at multi-tenancy as a key criteria for acquiring a new piece of functionality is basing their decision on factors that are not directly relevant to their TCO, all other factors being equal.

Phil Wainewright pokes back:

To say that multi-tenancy is only of interest to vendors and has no relevance to customers is a bit like saying Wintel compatibility is only of interest to PC manufacturers, and so customers should not worry about it. Try telling that to the enterprises that invested and wasted millions of dollars in rolling out DEC Rainbows or IBM PS/2 Microchannel machines in the mid- and late 1980s. If a vendor is selling you a proprietary dead-end that will be obsolete before its time, I’d say that’s a factor of huge importance to customers. Don’t let anyone tell you any different.

Phil is talking about the technical dead ends Phil believes are represented in private cloud explanations. Not satisfied with that explanation, Josh comes back with:

...single tenant vendors can be compatible with the promises of multi-tenancy in terms of updates, pricing, support, etc. even if they stick with their single tenant model, much like any vendor could build to the Wintel spec, and run Windows, without having to duplicate the IBM PC.

Not one to miss a good technical punch up, Bob Warfield dives deeper into the TCO and cost issues arguing that:

Multitenancy is a little different.  Instead of 10 copies of the OS, 10 copies of the DB, and 10 copies of the app, it has 1 OS, 1 DB, and 1 app on the server.  But, through judicious modifications to the app, it allows those 10 customers to all peacefully coexist within the app just as though they had it entirely to themselves.

Can you see the pros and cons of each?  Let’s start with cost.  Every SaaS vendor that has multitenancy crows about this, because its true.  Don’t believe me?  Plug in your VM software, go install Oracle 10 times across 10 different virtual machines.  Now add up how much disk space that uses, how much RAM it uses when all 10 are running, and so on.  This is before you’ve put a single byte of information into Oracle or even started up an app.  Compare that to having installed 1 copy of Oracle on a machine, but not putting any data into it.  Dang!  That VM has used up a heck of a lot of resources before I even get started!

If you don’t think that the overhead of 10 copies of the stack has an impact on TCO, you either have in mind a very interesting application + customer combination (some do exist, and I have written about them), or you just don’t understand.  10x the hardware to handle the “before you put in data” requirements are not cheap.  Whatever overhead is involved in making that more cumbersome to automate is not cheap.  Heck, 10x more Oracle licenses is very not cheap.  I know SaaS companies who complain their single biggest ops cost is their Oracle licenses.

Bob's post deserves more than I can reasonably give it. It makes a persuasive argument and is definitely worth the re-reading. However, there's a piece of this discussion that's missing in the context of enterprise applications.

The TCO default argument

Josh thinks that single tenant players can and will make a successful cost bid against multi-tenant players. I think he is wrong. Bob starts that discussion but let's look further. The real cost test for me comes in the history of SAP Business ByDesign. Much maligned for getting it wrong, SAP kicked off with the idea that no finance chief would ever want his data in a multi-tenant environment. So it invented 'mega-tenant' where customers sat on physically separate blades in the data center. To SAP it was entirely logical. To the rest of the world it was nutty. And so it proved.

SAP quickly discovered that the provisioning and maintenance costs of all these wee tenants meant its business model collapsed. Roll forward 18 months and SAP has re-engineered for multi-tenant. So in that sense you might argue that Josh is correct. Benefits to SAP. Well actually it's benefits to both SAP and customers. Faced with the mega-tenant problem BYD would have ended up costing customers a multiple than SAP had reckoned and publicly stated. Or BYD would have been killed stone dead as a cost sink. So yes, there is a clear cost advantage from which both customers and vendors benefit.

Multi-tenancy promises so much more and this is where the vendors have done a shocking job in getting it right, not helped by the muscular arguments that occur between strongly opinionated technically minded analysts. Let's go back to Workday.

During the briefing, Workday said inter alia that its current average current deal size is $1 million. That is far from shabby when you consider that it is going up against SAP, Oracle and the like. But also remember that's a per annum figure. With 160 customers it doesn't take a genius to work out that Workday has more than crossed the threshold for being considered a serious player in what is still a nascent market. We can confidently conclude that Workday has successfully pushed the TCO argument.

During his go to market strategy session, co-founder Aneel Bhusri said he plans to make multi-tenancy and in-memory database cornerstones of his marketing argument. Stan Swete, CTO picked up the beat by explaining how Workday leverages these ideas and a whole bunch of others to not only deliver lowered TCO but also new capabilities. Check the buzzword compliant slide (above) used as an introduction. Check Stan's cost take on the topic. Fast forward 24 hours and a session I had with SAP where they view Workday as a danger. Guess why? In part because of the benefits that accrue from Workday's technology play as a SaaS vendor.

But even then I don't think playing the tech card to that extent is terribly helpful. As I said at the time, the only people who really care are the tech analysts. Most business users I speak with could care less about multi-tenancy. They care about benefits. TCO is only one that can be well understood and has plenty of legs to it in both recessionary and growth periods. The distinctions are important but can be kept in background as vendors walk through the many business benefits. Longer term, TCO will become table stakes.

Beyond TCO: it's the data stoopid

In one of Josh's pieces Subraya Mallya says:

Most of you proposing the Multi-tenancy here are just talking about TCO as the benefit. But if you look at the rate at which Cloud Computing is bringing down the cost of infrastructure, going forward, the argument will be less of TCO. So what else are the benefits of Multi-Tenancy to the Customer?

One of the key drivers, for SaaS vendors, to go with Multi-Tenancy, besides economies of scale/TCO, is the potential pot of gold that gets accumulated with aggregate data across tenants. Conventional thought has been the data resides in a single db across tenant-owned-slices you could aggregate them and roll them up to create analytical models. Generating patterns, identifying trends and insights to drive high value analytics (descriptive & predictive) is going to be the future of Information Management. But that is now possible with massively parallel NoSQL technologies (Hadoop, MapReduce) even if data were to reside in multiple datastores.

Bingo. I have long argued that multi-tenant architectures offer transformational benefits from the ability to aggregate data. That is not possible in a single tenancy situation. Workday isn't ready to contemplate that notion just yet any more than any other vendor. Except perhaps NetSuite and a small handful of the very small business apps vendors that are building in these capabilities. I recall a conversation where it was said that NetSuite saw the recession coming before it became public knowledge by virtue of the transaction trends it saw in its customer portfolio. Try doing that from your single tenant application.

How about selling anonymized data to banks, other financial institutions, telcos, insurance companies, healthcare organizations...the list goes on. Why would they buy these data? Application tech vendors tend to specialize in certain business sectors. Multi-tenant SaaS vendors are collecting prime data that has genuine value third parties cannot get.

A bank may see ins and outs of bank accounts but those numbers are meaningless without context. A first step I am seeing is where some vendors are suggesting giving client bank managers shared access to aggregated financial data in real-time. Why? Because the bank can marry what they are told with what they see and so make much better informed lending decisions. Add in the ability to offer comparative trend data and you've got something extremely powerful. Is it a viable model for enterprise? That remains to be seen but there is no obvious technical barrier. But hang on - isn't this another tick in Josh's favor? Well yes, but not quite.

Long term, I believe the ability to powerfully slice, dice, form and reform data out of multi-tenant systems will become the place where customers see huge value that is way beyond TCO. If I can benchmark performance in real-time or spot trends and compare, again in real-time, then my ability to take corrective or revenue enhancing action is vastly improved. Do I maintain a competitive edge? Of course because it isn't the availability of data that matters but the ability to execute plans against what I am seeing.

Does this require careful handling to ensure that confidentiality is not breached? You betcha. But just as in the consumer world we have forgone privacy in the name of getting help from Google, there is no reason to believe the same wont hold true in the enterprise. The upside potential for benefit is simply too big to ignore.

Editorial standards