Interana's a SaaS, right? Nope. Here's why not.
What Interana has been so far is software that we deploy and manage in dedicated clusters on behalf of customers like Microsoft, Bing, Sonos and Comcast. Usually in their dedicated AWS or Azure cloud environments; sometimes on remotely managed hardware running on premises in their datacenter. These customers often have hundreds of users doing dozens or hundreds of complex behavioral queries per day on up to trillions of events coming from many different data sources. Our turnkey management ensures high availability while the dedicated deployment model ensures performance isolation and security. However, it does come with a price tag and requires customers to give us the administrative rights to manage their data clusters.
That's why we've been working hard this last year to make Interana available to a broader market than we can address with this fully managed offering (although we'll keep offering that, don't worry!)
So you might think we would make an on-demand SaaS version of our service. We're not. We're working on releasing Interana as downloadable software that customers can install and run themselves wherever they want - on prem or in any cloud. Sounds a little old school, right? We know.
Turns out that for interactive analytics, SaaS isn't quite the magic bullet that it is for something like billing or customer support software. Many benefits of SaaS are arguably temporary economic advantages that will disappear as soon as the SaaS providers have to get profitable. And there are increasingly obvious downsides and limitations of SaaS for what we do.
First off, you have to get the data to the SaaS. If the only data source you care about is client instrumentation, not a big deal. But that only gives you mobile and web analytics from the client perspective. Increasingly, our customers care about event data from pipelines and server applications running in their own environment - which for big customers is still often on premises. The few viable event data pure cloud plays like Sumologic, Wavefront and New Relic typically rely on some sort of on premises queuing and even data reduction to get data from on premises to the cloud - which works for their security or IT focused use cases but is a non-starter for Interana's business-focused exploratory analysis use cases that rely on near-real time feeds of 100% of raw data.
Then, you have to trust the SaaS with your data. A handful of leading SaaS analytics and log management providers have gone to great lengths to get PCI and other security and privacy compliance certifications. But most start off immature with data practices that would make you cringe. Buyer beware.
Then, the SaaS has to retain your data as long as you want or need it. This is where the economic question comes in. Overfunded SaaS players like Wavefront like to tout that you don't pay for storage, just ingest rates. We'll see how that survives future funding climates. More prudently managed smaller SaaS players severely limit your data retention - or offer products that just provide limited summary metrics so they don't need to keep the raw data. Extended retention times are at costly markup to the cloud storage costs these small SaaS players are paying - often far less economical than what their larger customer prospects can negotiate from the cloud providers directly.
These are reasons that a lot of potential customers for behavioral analytics have historically gone with their own roll-your-own data infrastructure based on open source technologies that they can deploy themselves, anywhere. Yet these approaches severely limit flexibility for end users, with most new analysis requiring pre-processing and data engineer or data scientist intervention.
So where we're headed is Interana software that is easy to download, install, deploy and run yourself wherever your data-generating services live today - whether in AWS, Azure, Rackspace, or on servers in your own datacenter - and that can be managed using your existing IT tools and processes. You add your own data and open it up to your users however you want. You'll be backed by our world class customer support as well as by the community of other Interana administrators and users on Interania.
We'll have a few challenges to address together - most notably the resource requirements of our intensive data analytics processing that mean your servers whether in the cloud or on premises will need to meet certain performance specs that are beyond that of typical enterprise applications. But we and the community will provide guidance on specs for common platforms as well as diagnostic tools.
If you still want us to manage it, we'll keep offering that option. And it's likely that we'll find top tier cloud partners that want to offer Interana-as-a-service that might make sense if your data lives in their services today.
But as a young software company with one bullet, we're going to use it to make great software that can run anywhere for anyone, not a limited service for a few.
Interested in running Interana for yourself? Apply to join our special Friends & Family Program to get an exclusive first look at what we're working on.