No Comments

Summary:

The enterprise is adopting disparate services and using them to build federated applications as opposed to deeply integrated and less flexible programs. Orchestrate thinks it has a solution for the database drama that can ensue.

tony_bw2

The enterprise software stack is changing. What used to be a monolithic, integrated platform is slowly breaking up into a variety of services delivered by companies that range from Salefsforce.com to Sendgrid. This concept of piecing together services to build federated apps is becoming accepted, but when you start thinking about getting and sharing data in such a fragmented world what looked cheap and simple starts getting difficult again.

This was the founding thesis of Orchestrate, a Portland, Ore., startup that is trying to build a service that doesn’t deliver one database as a service, but rather replaces a variety of databases delivered as a service through its API. The company launched today after several months in beta and a $3 million round of funding.

A company using Orchestrate writes its data to the Orchestrate API and it is stored in an index on Orchestrate’s Amazon Web Services servers. Then Orchestrate writes the data to a variety of database types where that data needs to live. It stores some of it in memory so the company can access it faster (the company can dictate what gets placed in memory). Antony Falco (pictured), the CEO of Orchestrate, says that the results so far are reads and writes that compare with any other cloud-based database service. Those services include Amazon’s Redshift, DynamoDB and RDS, Google’s Datastore and Heroku Postgres.

For those where database speed is essential and are currently running on bare metal, highly optimized products, this isn’t the service for you, but Falco is making the case that those use cases are becoming more rare in a world of federated software where agility trumps speed. Which bring up the next possible issue — that of consistency. Here’s where Falco has differentiated his company.

When you store data across many different servers around the world, as many companies are doing, it becomes more challenging to make sure that the results are up to date. For example, if someone updates customer data in China, the information needs to be the same if someone checks that same customer account in the United States a minute or seconds later. Falco says that to deal with this, he’s decided that Orchestrate buys into the concepts of an immutable object, saving only one version of the data, but labeling it so anyone grabbing the data knows what version it is. So if you save Version 5 of a file but someone is pulling down version 4, then that person can either accept that Version 4 is good enough or decide to wait until Version 5.

Falco says that in most cases developers don’t even want to think about this choice so Orchestrate just does this for them but flags that it happened. He may develop a feature later that lets developers get more manual in making the call on whether to wait or not, but again, he believes that for many people building these services good enough is good enough. It’s an interesting shift but one we’re seeing happen across a wide swath of technologies as we rely more on IP and distributed systems. Frequently, agility and the ability to move quickly in IT deployments are trumping the exacting and deeply integrated world of old. It may have performed better but that’s no longer the top concern at most businesses.

At least, that’s what Orchestrate is banking on with its launch.

Disclosure: Orchestrate.io is backed by True Ventures, a venture capital firm that is an investor in the parent company of GigaOM. Om Malik, founder of GigaOM, is also a venture partner at True.