Big Data

Attention developers with enterprise search problems: Qbox wants to talk to you

Arkansas startup Qbox provides managed cloud hosting for open source search engine Elasticsearch. Learn how Qbox serves its primary market and more in this Q&A with the CEO.

searchistock042114.jpg
Image: iStock/banarfilardhi

Qbox (formerly known as StackSearch) provides hosted Elasticsearch for its enterprise and smaller customers through dedicated clusters on Amazon's EC2, Rackspace, and IBM's SoftLayer in 16 data centers on five continents.

Elasticsearch, the open source project, is "a flexible and powerful open source, distributed, real-time search and analytics engine" that is built on Apache Lucene.

Netherlands-based Elasticsearch, the company, is the evangelizer and promoter behind the open source project, and is on a mission to "organize data and make it easily accessible" and "make real-time data exploration available to anyone."

TechRepublic recently held a Q&A session with Qbox CEO Mark Brandon, and talked about enterprise search, Qbox's goals and recent product offerings, its focus on developers, and the sheer pain of dealing with inadequate solutions in big data prehistory.

TechRepublic: What is your shortest, most effective description of Qbox?

Mark Brandon: We offer solutions, support, and hosting for Elasticsearch, for big data exploration and analytics.

TechRepublic: Please define the term enterprise search.

Mark Brandon: It's a very bastardized term, but I'll give you my definition. When a lot of people think of search they think of Google, a third party that crawls the web, indexes it, and makes it searchable. Enterprise search is when you need to crawl and index your own internal resources, make those searchable, and customize them for your own use case.

If an enterprise has 80,000 people and all kinds of data on them, you will want to make it available to your own people because they will want to know who is in the organization and who they can collaborate with. You don't want to make that available to the public — you just want to make it available to your own people. So essentially you have to create your own little Google for your own resource.

TechRepublic: Could you give a quick briefing on Elasticsearch?

Mark Brandon: They drive the product, and have a team of developers managing the open source project and making the platform better. From a business standpoint, what they focus on is training and evangelizing the platform, and professional support. The professional support could be for an on-premise version of Elasticsearch, or hosted versions as well.

What we focus on is that segment of users that need managed services. We do professional services, and hosting, and when we have an opportunity to sell professional support, we do that in partnership with Elasticsearch.

TechRepublic: Who is your primary market, and how do you serve them?

Mark Brandon: The person we target is really the developer who has an enterprise search problem. This can be across the board, from very small startups to Fortune 500 companies, because data sets are growing exponentially. All kinds of data sets are made available — they are expanding faster than the considerable improvements in computing power and storage. There might be a small startup that is developing an API to log events or search products or search large, open databases. Or it could be a Fortune 500 company that has tons and tons of retail analytics, or business events of all sorts.

TechRepublic: How was Qbox founded? What was your entrepreneurial idea?

Mark Brandon: We started out as an e-commerce product search company, focusing on making front-end product search fast and flexible for customers. Our product didn't get much traction. Part of that was because there is a very long sales cycle with e-commerce merchants.

Many of our customers asked whether they could use this for non-commerce activities, like retail analytics or other exploration tasks. Since we already built these massive clusters for Elasticsearch, we thought why don't we just make this available as a service, and see if that takes off. And it did start taking off. So before long we kind of shelved the e-commerce product and are now focused on the managed services.

TechRepublic: Last month you launched Qbox Enterprise. What goals are you pursuing with that offering?

Mark Brandon: It goes back to the whole dawn of cloud computing. There are those companies that don't ever want to use a third-party provider's infrastructure for whole plethora of valid reasons. It might be regulatory or security or whatever. So those customers were kind of shut off to us.

Qbox Enterprise is really just a means for us to serve those customers that don't want to use our infrastructure. So if they want to use their own cloud account for their own on-premise infrastructure, then we can work with them through a support contract that gives them our expertise on their infrastructure.

TechRepublic: What kind of traction are you getting on Qbox Enterprise?

Mark Brandon: I can't say off the top my head. I would say it's a much bigger decision for users, because it's a much higher dollar contract. We have a pipeline that were working with — it just takes a lot longer to start and close the contracts because it is a mid-five figure contract rather than a couple hundred dollars per month.

TechRepublic: I was amazed by the advances I was hearing about when I started covering enterprise solutions two years ago. I remember in a previous job beating my head on the wall trying to do things with Microsoft Excel.

Mark Brandon: One of the apocryphal definitions of big data is anything that crashes Excel. If you are on a standard laptop, Excel will start to choke when you get to about 70 to 80,000 lines. And you can put it on a relational database and that will start to choke at a few million lines. And at that point, you have to kind of cluster it or scale it and put a lot of computing resources behind it. And what if you have a billion lines? It's not really feasible to have hundreds and hundreds of servers running an Oracle database where the license is $25,000 per node. If you have 1 billion lines, you're looking at dozens of nodes. And the cost to get your insights that way is prohibitive for most companies. That's why it's so important.

TechRepublic: What are your immediate goals as a firm?

Mark Brandon: We have a whole lot of things on our roadmap. We are working on a lot of projects to enhance Elasticsearch ourselves, making it easier to deploy and manage, or for specific use cases, like e-commerce. We might, for example, dust off that e-commerce application that we started off with, now we have the resources to make it better and be better able to sell it and market it.

About Brian Taylor

Brian Taylor is a contributing writer for TechRepublic. He covers the tech trends, solutions, risks, and research that IT leaders need to know about, from startups to the enterprise. Technology is creating a new world, and he loves to report on it.

Editor's Picks

Free Newsletters, In your Inbox