After years of selling backup software, Infrascale Inc. is now on a path to de-emphasize pure backup in favor of...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
disaster recovery as a service.
In 2015, Infrascale launched its Cloud Failover Appliance (CFA), a 1U hybrid box that offers guaranteed recovery of applications, physical servers and virtual machines in 15 minutes or less. The CFA hardware is based on technology Infrascale picked up through its acquisition of Eversync Solutions in 2014.
CEO Ken Shaw said enterprises using Infrascale backup are getting more sophisticated in their approach to data protection, choosing continuous availability of applications and data over traditional backup. Evolving customer expectations prompted Infrascale to shift gears to provide on-demand disaster recovery. Shaw said backup and disaster recovery (DR) are converging, manifested by "the emergence of DRaaS as a product category."
We spoke with Shaw about the evolution of Infrascale backup technologies, changing customer expectations and his company's attempt to carve a place in the crowded DRaaS market.
How many Infrascale customers buy your backup, and how many customers are buying on-demand DRaaS?
Shaw: Let me start by saying what we aren't about here at Infrascale. Backup, as it turns out, is really lousy at disaster recovery. We are using the cloud to disrupt DR. We can provide cloud failover with 15-minute guaranteed recovery at about 10% of the price of the major incumbents. We don't just give you back files or folders, the way a backup tool does. We can give you back an entire running data center, with applications, data and servers that are interconnected, internetworked and interoperating.
Why did you shift the focus from Infrascale backup to DRaaS? Do customers understand the distinction?
Shaw: The problem with traditional disaster recovery is that it's based on [use of] reserved infrastructure. I'm on a bit of a warpath against that approach. If I have 50 Oracle boxes and want to set up a DR site, it means I have to go out and buy another 50 Oracle boxes, rack and stack them together ... and triple my production costs to get a fully redundant system. That approach is just crazy. Our Infrascale cloud failover can achieve the same outcome on the fly. We only run infrastructure when an actual disaster has taken place, so we can provide service at a lower cost.
How do you position Infrascale DRaaS to your backup customers?
Shaw: We still bundle all our backup technologies for free. We have tens of thousands of original customers running the Infrascale backup-only legacy platform and are in the process of up-selling them to DRaaS. It's all one big product platform, with different modules that can be turned on and off.
What happened to the Infrascale Cloud Backup Accelerator (CBA) hardware you rolled out in 2014?
Shaw: We rolled the CBA technology into the Cloud Failover Appliance. The CBA hardware is now the smallest model in the CFA lineup, which scales from 4 TB on the low end to 400 TB on the high end. It scales out, so you can chain three 400 TB appliances together to build a petabyte environment.
The 15-minute service-level guarantee is aggressive. How often has Infrascale failed to meet it? And what happens when you can't meet the DRaaS service-level agreement (SLA)?
Shaw: We have never failed to meet it. We've built in plenty of buffer for ourselves. Each individual server will take between 30 and 90 seconds to fail over. We run everything in parallel, so our DRaaS can bring up 200 servers in parallel. Because we use the cloud, we can do DR on demand and have a lower cost structure. There's a lot of intellectual property that goes into being able to spin stuff up quickly, do the driver injection, translation, rehydration and deduplication. We avoid the need for using a second [physical] site.
If we fail to meet the guaranteed recovery SLA, we would refund the customer's money or apply the refund to their ongoing service contract. So far, that hasn't happened.
Which cloud do most customers prefer: public, private or Infrascale cloud?
Shaw: Half of our deals go in the Infrascale cloud, and that's because we can get aggressive on pricing versus the public cloud providers. But we want to avoid religious warfare around private or public cloud. We support Amazon, Microsoft Azure, IBM SmartCloud, and we'll support Google Cloud Platform. We also run on OpenStack, VMware and Microsoft Hyper-V. We basically have an option for everyone. The most popular option by far is our cloud. The second most popular option would be Amazon, and next would be customers that use their own private cloud.
How is your DRaaS priced?
Shaw: Our base pricing is $190 per terabyte, per month for recovery of up to 200 servers. That includes the necessary equipment for your environment. We figure out which hardware you need to run on premises for local storage and install it as part of the package. We are trying to radically streamline and simplify that process [as a way] to get market share.
Ken ShawCEO, Infrascale
How much traction is your DRaaS model getting with enterprise customers?
Shaw: Without getting into specific numbers, I can tell you that adding cloud failover as a service is a big part of our revenue growth in the enterprise business-class segment. We launched 15-minute guaranteed DRaaS in 2014 and tripled our sales from 2014-15, and we're on track to triple sales again in 2015-16.
Describe the role of embedded WAN optimization in your CFA hardware.
Shaw: That is a critical piece for building scalable DRaaS. WAN optimization enables replication that is four to five times faster than it would be otherwise. You need that kind of replication speed to make cloud-based DR a reality.
Think back to the enterprise customer relationship management market in 2000. It was enterprise software that only the biggest companies could afford, took a long time to implement and then everyone hated the outcome. That's kind of what the enterprise DR landscape looks like today. Then, Salesforce.com came along with a cloud-based, on-demand, deploy-tomorrow model. We want to do for DR what Salesforce did for CRM.
What is next on Infrascale's product roadmap?
Shaw: Right now, we support several different Linux platforms and plan to add support for more Linux systems this year. We think we can be aggressive and get the guaranteed recovery time even lower than 15 minutes. We're also planning to build sophisticated workflow tools that make it easier for customers to set up all the orchestration themselves. We're trying to make it so customers don't need professional services to make DRaaS work.
Explore rise of DRaaS, key matters to consider
Make the right DRaaS vendor choice
Investigate cloud DR strengths, weaknesses