Tell us a bit about how Delphix came about.
I had founded Avamar in 1999, doing data deduplication. We were ahead of the market but we sold it prematurely [to EMC] for $165m, just before the market really took off. But while I was at EMC I realised that the only way you were going to solve the challenge around databases was to build a specific product for that. Delphix doesn’t do deduplication, what we do is prevent duplications from happening in the first place.
Tell us a bit more about how it works and what customers use it for?
Companies use it for a couple of areas. They need it for dev and test, and also when they are moving data for analytics or data warehousing. Our software engines basically collect data from the databases and record all changes to the databases. Then we deliver it to the right team at the right time. Applications reside across multiple data centres, so you drop our software in wherever the data resides.
What sort of benefits do customers typically get?
We’ve seen release cycles reduced by 20-80%. Companies spend all this money on R&D but often developers are sat waiting for data. They’re basically haemorrhaging money. Delphix gets the data to the right place faster – on-demand, on time.
Is your software client-server or delivered via the cloud?
Content from our partners
We deliver it as a virtual machine, so it’s cloud ready. Companies choose whether to run it in their private cloud or in a public cloud hosted by one of our partners.
I know companies are usually very wary of adding anything to their mission critical, production databases. Does your software need agents to sit on a database and if so is there any disruption?
We spent a lot of time designing it to be as non-disruptive as possible. We use existing database vendor APIs, for example the Oracle API. That’s so it doesn’t add complexity or risk to production systems. Data is being moved out of databases all the time, for backups, BI and reporting, and for application teams to do unit testing, QA testing. Dropping Delphix in radically reduces the data being constantly pulled off the database, so it actually reduces the load on the production database.
Which databases do you support – all the majors like DB2, Oracle, Microsoft?
Today we support Oracle, Exadata and Microsoft.
No DB2? Will you add others?
We are in development for Postgres, MySQL and Sybase. We may do a DB2 version but it’s not in development yet.
The other three you mention, will they be available in the next 12 months?
Yes.
What about the company itself, how many customers do you have?
We have about 170. 50 of the Fortune 500 and 100 of the Fortune 1000. We’ve seen average growth in the past few years of 300%, this year so far we’ve already grown over 200%. We did $30m revenue in the last four quarters.
Are you profitable?
We’ve been cash flow positive for the last couple of quarters. It depends where we are investing in staff.
Is it too early to be thinking about an exit strategy – an IPO or trade sale?
We have a non-exit exit strategy. Meaning that we want to build a standalone company and an IPO is just a point in time along a continuum. So it’s just a question of when is the right time for an IPO.
Where would you describe the market – do people realise they need your technology yet?
I’d say we’re pretty early in the adoption cycle. Today we have no direct competitors. It’s a very complex solution to build because databases are all different and there are lots of interfaces. We’ve got patents around our algorithms and there are a lot of checkpoints you have to get right – file systems, performance requirements, in-memory virtualisation of data blocks. There are several bridges to cross so we’ve got quite a lead.
I notice you have a product aimed specifically at SAP applications. Is there any implication for SAP’s HANA customers?
We have lots of customers who use SAP HANA and they use Delphix to feed data into HANA for further data warehousing or analytics. Offloading the work to Delphic means you get the information in HANA fresher and faster.
