The service delivery platform market is going to see disruptive players in 2012. One of these offerings is Restcomm that offers an open source Telco PaaS solution that will get to version 1.0 in March 2012. I interviewed the director of Cloud Engineering of Telestax, Thomas Quintana, who is responsible for driving Restcomm within the Mobicents community.
What is RestComm?
RestComm is what I like to call a “Web Driven Communications Platform”. In order to understand what I mean it’s useful to have a basic understanding of how RestComm works. When RestComm receives a phone call for example, it calls out to some resource on the web to instruct it on how to handle a call. In turn the resource itself can turn over control of the call to other resources on the web. Hence my description of RestComm.
Who is the target user?
In developed markets that is an interesting question because the target user has always been web developers but we have noticed that the RestComm instruction set maps nicely as a DSL (Domain Specific Language) on top of development platforms such as Java, Ruby, Python, etc. This in turn has made mobile applications developers and desktop applications developers target user communities as well.
In emerging markets the target users are operators who want to deploy a PaaS (Platform as a Service) and provide services to its customers and SME’s
What is the RestComm Vision?
The RestComm vision is to offer a 100% open source and service provider agnostic web driven communication platform to the open source community. Woah, that’s a mouth full 🙂
What is on the direct roadmap for RestComm?
There are a quite a few goals on the RestComm road map but I will list the ones I think are currently most important:
- 100% Twilio API compatibility (Available as of ALPHA 2 which will debut in a few of weeks)
- 100% Tropo API compatibility
- Monitoring support
- Web based management interface for easy management
- Support for other communications networks such as Skype and Google Talk
Yes, as a matter of fact most if not all of the Twilio wrappers will work with RestComm with very little effort (possible as of ALPHA 2). Usually all it takes is just pointing the API at your RestComm server.
When will release 1.0 be available?
The RestComm FINAL release is scheduled to be available in March.
What will be in release 1.0 that is currently not there yet?
We are doing a lot of work on RestComm based on community feedback from early adopters so it’s hard to tell everything that will make it in to the FINAL release but the following are things that will definitely be present.
- A super set of the Twilio API to support functionality currently not covered by the Twilio API but requested by our development community (e.g. faxing, call leg and conference room recording, etc.).
- JDBC and MongoDB support (plus it’s very easy to add additional back-ends)
- Support for any SIP origination/termination provider
- Support for more international SMS aggregators
- High Availability support
- and many more things stay tuned 😉
Below are two features that we would like to have available by the FINAL release but are still investigating.
- RestComm application Fail-Over support
- Very large self organizing conference rooms (1000+ participants)
How will RestComm scale?
This question is a bit difficult to answer because RestComm’s ability to scale is limited only by the infrastructure that it runs on. In order to answer this question I would like us to assume that most of the RestComm deployments will be on private/public clouds with sufficient resources. Now that we have some context RestComm scales horizontally on two layers the RestComm interpreter itself and the media gateways. Each RestComm interpreter can drive tens to hundreds of media gateways providing scaling for media and by placing SIP load-balancers in front of the RestComm applications servers the interpreter itself can be scaled horizontally. We are currently investigation media gateway fail-over where if a RestComm interpreter becomes unavailable the other interpreters in the cluster will take over its media servers and continue executing the RestComm applications.
In order to provide a demonstration on how to accomplish large scale scaling the Mobicents team will publish a paper on how to deploy RestComm on the Amazon EC2 cloud with hundreds of instances once we reach our BETA release in February.
Will there be a GUI management interface?
Yes, we are planning on releasing a web based management interface but we are currently focusing on providing a stable release.
Will there be a monitoring interface?
JMX and SNMP support are currently available for the Mobicents Sip Servlet Container which RestComm sits on and that is how we monitor RestComm for now. In the future RestComm will provide its own monitoring interfaces.
Will Telestax offer support/SLAs for RestComm?
TeleStax enables Telecommunication Service Providers and Enterprises to create scalable communication applications based on Open Source and Open Standards. Therefore, TeleStax will provide supported versions of the “RestComm Platform for Production” aka RCPP (like JBCP) to its end customers. The support subscription will be for developement as well as production support with flexible SLA’s to suit the requirement’s and demand’s of end customers. For further details contact TeleStax –http://telestax.com/contactus/
In recent weeks I had the pleasure to talk to the team behind Mobicents. I have been pleasantly surprised by Restcomm. Restcomm is Twilio for the Private Cloud. Telco 2.0 SaaS for private cloud. Tropo APIs are also on the roadmap. Mobicents is starting a revolution by moving away from telecom standards and moving to the new Cloud telecom standards. Telecom engineers are no longer needed to make enticing value-added services. Any web developer can make telecom apps in minutes and integrate them with Web 2.0 and social networks.
Is Restcomm a threat for Twilio? Quite the contrary. Many larger companies did not want to move to Amazon because of fears of vendor lock-in. Eucalyptus brought a way for public cloud apps to move to a private cloud. Restcomm will allow companies to move their telco apps to the private cloud when they become a large hit. Developers could even start from a private cloud deployment and move apps to Twilio when spikes in demand happen, a.k.a. cloud bursting. In general Twilio is very likely going to get more customers instead of less by having a valid open source alternative.
Mobicents is also undergoing large changes. There has been a shift in direction at Red Hat and the Mobicents team started their own company called Telestax. The company is independent from Red Hat, however it will partner with Red Hat for telecom opportunities involving RHEL and JBoss products.
Too many people in the telecom industry are still discussing which API is the best: Parlay, JAIN SLEE, Sip Servlets, GSMA OpenAPI, etc.
However even these APIs are too complex for some people. In this case you can use graphical drag-and-drop environments like for instance QuickFuseApps. You can also opt for flash modules that give you all the functionality you need. Ribbit has some nice ones.
Also on the phone side, drag-and-drop is coming on strong. Google´s App Inventor for Android is a good example.
What does this mean? More and more developers and end-users will be able to create Net Apps themselves. These Net Apps will quickly become complex applications that will often bridge the gap between mobile devices and server & cloud solutions. They will very likely also span every aspect of daily life, e.g. social networking, business, entertainment, etc.
What does this mean for an operator? All the effort that is now put into creating attractive services will no longer be useful. A one-hundred person marketing team can not launch more and better services than a one-million net app creators community. So instead of focusing on finding and developing the next killer app, operators should focus on two aspects:
- Making sure all the building blocks are in place for the net app creators community to be productive.
- Connecting end-users with the proceedings of the net app creators community. In other words: make sure people find the right net apps.
The stakes are high because this is a winner takes-it-all game. Speed, easy of use, direct community feedback will be key. What are you waiting for?