Working at Intel certainly has its benefits, especially when it means working alongside the Chair of Open Source MANO's Technical Steering Committee, Adrian Hoban. A few weeks back, I had the pleasure of catching up with Adrian after his recent trip to OSM#3 Plenary meeting in Palo Alto, California. During OSM#3, the technical community met face to face to discuss architectural principles, alignment with ETSI NFV, status updates on the features of Release TWO, and to collectively agree on the goals and themes for Release THREE.

With Release TWO on our doorstep and a release date of April 27th, I hastened to ask Adrian about the key deliverables and what we could expect from the upcoming new code base. Adrian summed it up quite nicely: “Release TWO has advanced sufficiently that it is ready for operators to begin their RFx processes and field trials. It brings significant improvements in terms of interoperability, packaging & distribution, usability, data plane configuration and dynamic service assurance.”
I’ve captured some of the finer details of our conversation below under six separate value vectors:

1. Interoperability

Interoperability is a key strategy for OSM which ultimately will achieve three critical objectives for end users as network transformation accelerates:

1. Significant reductions in cost and complexity.
2. Flexibility to implement “best of breed” deployments.
3. Ability to leverage existing investments in their networks.

Several developments will be part of Release TWO that further accelerates this strategy:

• The code base now offers multi-disk support. This is important as many applications need to use more than one disk to operate effectively.
• The addition of OpenStack v3 APIs enables greater flexibility in an OpenStack deployment.
• An additional 15 VNF vendors were on-boarded during the ETSI Plugtests which is a testament of the ease of use of OSM’s Data Model and on-boarding process.
• Connector improvements in VMware vCloud Director which allows for greater interoperability between VMware vCloud and the MANO stack.
• Interop with public clouds like AWS which opens the door to automating NFV deployments in public clouds or in hybrid multi-site scenarios.
• Two new OSM Remote Labs provided by Wind River and VMware joined the network and substantially increased the number of combinations of VIMs and SDN Controllers available for the OSM CI/CD.

2. Packaging, Distribution and Installation

How easily you can install a system has a direct correlation to OPEX savings, not only from the point of view of time to setup but also from the angle of maintaining skilled in-house expertise and physical resources for the installation process. Simplicity is key to cost efficiencies and to the advancement of the network. Some of the developments in Release TWO include:

• A new package management update for the UI.
• A 50% reduction in the OSM memory footprint.
• Docker container images can now be used as a distribution model.
• UI Composer now has a dynamic cross-reference drop-down list to ensure the validity of Data Model entries.

3. Usability

Usability is equally fundamental to realising savings in OPEX, not only through the ease of use but also through improved fault detection and recoverability. The advancements in this area include:

• Access to VNF consoles through the User Interface (UI).
• Support for cloud-init.
• Common Logging & Exception Handling with the alignment of OpenVIM logs and exception handling within the OSM framework.
• Availability of new OSM Remote Labs makes it very easy to use a variety of test infrastructures.

4. Security

It’s quite common when discussing security that the focus is often on locking down parts of the network or limiting access. However in an NFV context, often the application requires access to all of the traffic to help inform or make decisions. The ability to disable filtering on ports has been added in Release TWO to give the necessary level of control to the admins.

5. Data Plane Configuration

Data Plane configuration has been extended to facilitate complex network deployments that are aligned with some of the needs of 5G infrastructure. Two new features include:

• A new Underlay Network Management feature is introduced which allows direct management of the SDN Controllers like ODL and ONOS.
• OpenVIM can now manage compute nodes running OVS. Previously, only Linux bridge was supported.

6. Service Assurance

The ability to guarantee network services to your customers is essential and not only in terms of uptime but also in the ability to grow and shrink network services on demand. Release TWO delivers an experimental Network Services Scaling feature allowing you to dynamically add or remove VNF instances from running services.

During our conversation, Adrian also added that aside from the technical advances during the Release TWO development cycle, the OSM Technical Steering Committee also engaged extensively with ETSI NFV providing feedback on the Information Model (IM) for the VNF and Network Service Descriptors. OSM was also a successful participant in the first NFV ETSI Plugtests earlier this year.

In response to a final question about OSM’s plans beyond Release TWO, Adrian concluded that Release THREE is focused on delivering a production-ready solution and is expected to bring further enhancements in the areas of service assurance, usability, resiliency and security. Full details of Release THREE will be locked down in the coming weeks and will be announced shortly on the ETSI OSM portal. 

A parting message to the OSM community is to stay tuned for the exciting new upcoming developments. Thank you, Adrian, for your updates, insight and time.