Database-as-a-service vendor Orchestrate, which cut its teeth with developers and startups, today is rolling out Orchestrate Enterprise in response to the need to manage multiple databases on a massive scale. As companies increasingly develop applications through APIs and the Cloud, Orchestrate manages multiple databases, including Elasticsearch, Hbase and MySQL, in a single footprint while addressing new requirements for the Internet of Things (IoT) applications and Big Data.
Market analysis firm 451 Research has projected that DBaaS revenues will grow from $150 million in 2012 to $1.8 billion by 2016.
Yet says Orchestrate co-founder and CEO Antony Falco: “Unlike other companies, we’re not building databases, we’re building the value-adds on open source databases.”
Earlier this month the Portland, Ore.-based startup launched geospatial search to its multi-tenant platform, adding to Lucene search, time-series events and graph query types. It’s offering JSON Document DB to store and retrieve JSON objects in key/value collections and offering deployments on either a public or private cloud. It’s hosted on Amazon Web Services and testing new services on CenturyLink.
Basho veteran Falco funded the company with a $3 million seed round in April 2013 from True Ventures, Frontline Ventures and Resonant Venture Partners. It did a developer launch in February – it signed up more than 5,000 developers on the GitHub Student Developer Pack – and has spent the past nine months as a learning experience working with developers and a few customers.
The company grew out of demand for different query types, the need for a service that’s easy to use and that can scale to handle a lot of traffic, Falco said.
The company had to work on gaining stability with scale.
“Internet of Things uses cases have really ballooned up, and we’ve learned some valuable lessons there about the massive datasets,” Falco said.
“Having all the queries in one place is very important because it becomes very expensive to move big datasets from one database to another to do things like geospatial tagging. IoT is coming up very quickly in enterprises. Enterprises want to invest in scalable platforms, they want to be sure that investment in a technology today that will not be redundant or at end of life in less than three years, so continuing to add queries such as SQL is important.”
Here’s an example of the type of queries companies are asking for, he said: Show me everybody from California who has clicked “opt in” on newsletters with these keywords.
“So step one, stabilize the system for a big scalable multitenant cluster, then start adding the fun, juicy features that developers want,” he said.
Making it enterprise ready had less to do with technology changes to the product than with being able to provide assurances that enterprises want: ensuring the company could meet SLA requirements, documenting and more effectively communicating its security practices and being able to address total cost of ownership concerns.
One of the company’s investors provided an analyst to do a TCO analysis, Falco said, and determined that running a single query type on Orchestrate costs 46 percent less than an NOSQL alternative such as MongoDB, and 75 percent less for applications that require multiple models, such as search, time-series and geospatial, when the cost of ongoing management and support staff are factored in.
Those TCO figures can be a selling point with IT executives, Falco said, since many companies have architects who’d rather build their own systems. In the market, though, Orchestrate’s competing with Microsoft’s DocumentDB and Amazon’s DynamoDB.
Falco has big plans to continue building out the company, but tech industry analyst Curt Monash notes that there’s a poor track record for companies working to get software layers to talk transparently to a variety of data stores.
“Very few users are or should be comfortable writing any serious applications to an API that only is supported by one SaaS vendor. That’s why DBaaS offerings consistently flounder, ” he said. “Of course, the problem is greatly magnified when the vendor is a small and unestablished one.”
User Adron Hall, however, says so far Orchestrate “has hit the nail on the head.”
His company, Deconstructed, a platform for unifying customer data across devices and services, uses Orchestrate on the back end.
“Having [the various data stores] in one place where I don’t have to set up the databases, configure and maintain them is huge. It’s a massive, massive time save,” said Hall, the company’s founder and CEO.
He just recommended it to a developer building a minimally viable product that would require him to spend a few days setting up a key/value store and a graph store with some geospatial qualities for a product that is just a demo.
He believes Orchestrate will be more successful than previous attempts at combining datasets because it’s bridging disparate data rather than trying to combine databases into one, and doing so at a higher abstraction level.
Going forward, Orchestrate will focus on adding data centers and networks, with a European data center to launch in November; adding more query features such as faceting; introducing analytics, and offering multi-data center replication, allowing customers to segment their data and decide where it’s stored in order to speed up response times.
Orchestrate’s new price structure includes a free tier of up to 50,000 API calls per month for a single application up to custom enterprise deployments on-premise or with dedicated clusters with unlimited API calls.
