When is a Cloud Not a Cloud? When it’s a Single-Tenant Hosted Product

When is a Cloud Not a Cloud? When it’s a Single-Tenant Hosted Product

If you’re looking for a solution that offers the key benefits of a modern SaaS product, hold out for a provider whose underlying architectural model offers the benefits of a true cloud offering.

Charlie Oppenheimer is CEO of Loggly.

Not all “clouds” are created equal – or considered clouds at all, for that matter. With all due respect, single-tenant hosted products are one such instance. Just because a traditional software product is hosted by a vendor doesn’t make it the equivalent of SaaS.  Let’s face it – it’s not uncommon for successful licensed software companies that focus on operational intelligence or enterprise compliance and security to zig and zag as they evolve their business models to the cloud. Neither is it uncommon for them to maximize their best attributes in their marketing materials.

The difference between SaaS and a single-tenant hosted software “cloud”, however, is an important distinction. If you’re looking for a solution that offers the key benefits of a modern SaaS product, hold out for a provider whose underlying architectural model offers the benefits of a true cloud offering. And while your first reaction might be, “Who cares? Hosted software seems like SaaS as far as the user is concerned.” But here are the three reasons why customers should care about their “cloud” provider’s underlying model.

Single-Tenant Architecture Can’t Adapt at SaaS Speed

One of the biggest advantages a SaaS company has over a licensed software company is that development cycles are much faster and run in a tight, iterative feedback loop with customers. While licensed software products tend to deploy new versions on an annual or perhaps twice-a-year basis, a SaaS company deploys new “versions” constantly with new iterations every few days.

When a SaaS company deploys new capabilities, they’re measuring the results in real-time. They can see what’s working or what might be a bit confusing, watch overall usage patterns and performance, and rapidly adjust to that feedback. In fact, A/B testing of alternative implementations is a foundation strategy employed by SaaS companies to allow risk taking and faster innovation.  If a certain improvement doesn’t immediately work out as hoped, then a new iteration can be quickly deployed. .

In contrast, a single-tenant hosted product like Splunk Cloud can’t deploy new capabilities any faster than the pace of the licensed software product because the software and hosted offerings are the same. Software customers can’t possibly handle new versions every few days because of the operational overhead. They don’t want to have to have operational resources constantly running staging versions of new versions to test the next release and then deploy to production and start again. And this forces licensed software companies to batch up changes and slow the pace of innovation to the rate at which their customers can migrate.

SaaS companies are built to constantly stream new software into products with no interruption or overhead to customers. A new feature can appear on a product page without a customer even having to think about it. Obviously, bigger changes are noticeable (by design); but even then they do, everything possible to avoid requiring the customer to change how they work unless they want to.

Single-Tenant Architecture Limits User Behavior Analysis

There is a related and subtle problem for merely hosting on-premise software. SaaS products are built with the assumption that every aspect of the product’s usage and interactions are measured constantly and in aggregate so that anonymity is preserved. These measurements are monitored on scores of dashboards, alerts and ad hoc analyses.

But, licensed software products are not built with this kind of monitoring because the customers won’t have it. Does a F500 company license software with the idea that everything they do is reported back to the vendor? Of course not. Sure, certain high-level summary statistics can be reported back, but there’s no good way to push back low-level details anonymously because they come from a single uniquely identified source. When single-tenant architectures are hosted, they have the same problem as licensed software products.

Again, SaaS products are multi-tenant by design, so the dashboards and analyses are measured at a cluster level, aggregating many customers without having to watch what individual customers are doing.

Single-Tenant Architectures Limit Economies of Scale

There are other problems with hosting on-premise software. For example, from a capacity management standpoint, single-tenancy means that the amount of capacity allocated to that tenant is all that is available and any extra capacity made availabileto that one customer must be paid for.

On the other hand, a multi-tenant SaaS product enjoys the benefits of the law of averages. They can allocate far more capacity than any individual customer would need while amortizing the cost over all of them. At any point in time, some customers are using less than their subscription capacity, while others are using more. But through years of experience, they can plan capacity so that individual customer spike requirements can be absorbed easily and cost efficiently.

The SaaS Differences Add Up

A single-tenant hosted product will never match the advantages of a modern multi-tenant SaaS product. The economic advantages offer advantages to customers and the vendor as well.  Perhaps more importantly, a product designed to be licensed and used by a single company can’t offer the key benefits of fast, iterative development, behavior analysis and scalability that buyers are increasingly coming to expect. That doesn’t mean there isn’t room for both in this world – it simply means that an on-premise company delivers a different value proposition than a SaaS product does. It’s important to know which one you need.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish