Explore topic-wise InterviewSolutions in .

OPENCONTRAIL is an INDEPENDENT PROJECT that WORKS with other orchestrators.

OpenContrail is an independent project that works with other orchestrators.

There are DEPLOYMENTS BASED on JUNO.

There are deployments based on Juno.

We can deal with PARTITION of multi-facility DEPLOYMENTS ACROSS a WAN with a design in each local compute NODE’s peer with local control-nodes. This is based on the fact that control nodes peer with BGP.

We can deal with partition of multi-facility deployments across a WAN with a design in each local compute node’s peer with local control-nodes. This is based on the fact that control nodes peer with BGP.

Juniper can provide configuration guides for deployments with Juniper NETWORK EQUIPMENT; it is ALSO how we validate the solution. So it is EASIER to guarantee that it WORKS.

Juniper can provide configuration guides for deployments with Juniper Network equipment; it is also how we validate the solution. So it is easier to guarantee that it works.

No. SEVERAL production deployments USE VMs to run the CONTROLLER components.

No. Several production deployments use VMs to run the controller components.

 No. We CURRENTLY USE an open source PROJECT (IROND).

 No. We currently use an open source project (irond).

It defaults to using MPLS-over-UDP as encapsulation for IP TRAFFIC; it ALSO supports MPLS-over-GRE and VXLAN>.

It defaults to using MPLS-over-UDP as encapsulation for IP traffic; it also supports MPLS-over-GRE and VXLAN>.

No. It only HOLDS ROUTES for floating-ip ADDRESSES.

No. It only holds routes for floating-ip addresses.

Mirantis 6.0 does not support the upgrade of OpenStack itself, but it does upgrade the Fuel Master. An upgraded Fuel Master can DEPLOY new environments (6.0) while continuing to manage OLDER environments such as 5.1. We will be releasing functionality to deliver Mirantis OpenStack patches from a Mirantis-hosted repository in 6.1.

We will also provide an upgrade GUIDE for in-place upgrade of environments in the 6.1 timeframe. The upgrade guide is limited and will not cover all deployment configurations. We are taking a careful, pragmatic approach to document a process for upgrade, test it, and IDENTIFY gaps. In our 7.0 RELEASE, we plan to expand the coverage of upgrade and deliver tooling to support the process. Upgrade of Contrail-based deployments will not be supported by the upgrade guide.

Mirantis 6.0 does not support the upgrade of OpenStack itself, but it does upgrade the Fuel Master. An upgraded Fuel Master can deploy new environments (6.0) while continuing to manage older environments such as 5.1. We will be releasing functionality to deliver Mirantis OpenStack patches from a Mirantis-hosted repository in 6.1.

We will also provide an upgrade guide for in-place upgrade of environments in the 6.1 timeframe. The upgrade guide is limited and will not cover all deployment configurations. We are taking a careful, pragmatic approach to document a process for upgrade, test it, and identify gaps. In our 7.0 release, we plan to expand the coverage of upgrade and deliver tooling to support the process. Upgrade of Contrail-based deployments will not be supported by the upgrade guide.

VXLAN is SUPPORTED with a EVPN CONTROL PLANE.

vxlan is supported with a EVPN control plane.

We developed internal runbooks for integrating earlier versions of Contrail/OpenContrail components to MIRANTIS OPENSTACK 5 (Icehouse), but these are now DEPRECATED and not SUPPORTED.

We developed internal runbooks for integrating earlier versions of Contrail/OpenContrail components to Mirantis OpenStack 5 (Icehouse), but these are now deprecated and not supported.

Dynamic routing and service chaining. We ALSO believe that it is a more scalable implementation when COMPARED to the REFERENCE implementation.

Dynamic routing and service chaining. We also believe that it is a more scalable implementation when compared to the reference implementation.

There is a software gateway IMPLEMENTATION that is provided as part of Contrail. It BECOMES more efficient to use HARDWARE appliances once you cross a certain threshold of external TRAFFIC (e.g. 10G).

There is a software gateway implementation that is provided as part of Contrail. It becomes more efficient to use hardware appliances once you cross a certain threshold of external traffic (e.g. 10G).

YES. Most of the CONTRAIL PRODUCTION DEPLOYMENTS do this.

Yes. Most of the Contrail production deployments do this.

YES, it can. It doesn’t DEPEND on OPENSTACK. It is a SUPPORTED plugin in CloudStack, for instance.

Yes, it can. It doesn’t depend on OpenStack. It is a supported plugin in CloudStack, for instance.

YES, it is possible WITHIN REASONABLE LATENCY INTERVALS.

Yes, it is possible within reasonable latency intervals.

DEVELOPMENT BRANCH has MERGED into MAINLINE.

Development branch has merged into mainline.

Support for MULTIPLE L3 INTERFACES (as OPPOSED to LACP bond pairs) is PLANNED.

Support for multiple L3 interfaces (as opposed to LACP bond pairs) is planned.

The OpenContrail Fuel PLUGIN is scheduled for RELEASE with Mirantis OPENSTACK 6.1.

The OpenContrail Fuel plugin is scheduled for release with Mirantis OpenStack 6.1.

 OpenStack + CONTRAIL is being used by several carriers to deploy NFV applications. It allows NFVs to be deployed and processes IP traffic that TRANSITS between EXTERNAL NETWORKS.

 OpenStack + Contrail is being used by several carriers to deploy NFV applications. It allows NFVs to be deployed and processes IP traffic that transits between external networks.

Contrail INCLUDES a Neutron plugin. There are DEPENDENCIES when the API to that Neutron plugin CHANGES in a non-backwards-compatible way.

Contrail includes a Neutron plugin. There are dependencies when the API to that Neutron plugin changes in a non-backwards-compatible way.

Not YET

Not yet

Contrail implements the Neutron API. It can be seen as an ALTERNATIVE to the REFERENCE IMPLEMENTATION of Neutron through OVS.

Contrail implements the Neutron API. It can be seen as an alternative to the reference implementation of Neutron through OVS.

OpenContrail is an APACHE v2 licensed PROJECT. JUNIPER Networks OFFERS commercial support for OpenContrail.

OpenContrail is an Apache v2 licensed project. Juniper Networks offers commercial support for OpenContrail.

TYPICALLY via a tool such as HEAT that is doing APPLICATION MANAGEMENT. There are other open source and COMMERCIAL alternatives also.

Typically via a tool such as Heat that is doing application management. There are other open source and commercial alternatives also.

Currently most DEPLOYMENTS use Horizon as a TENANT UI and the Contrail UI as a network admin/secops UI. There are Horizon extensions for API extensions, but admin-centric API extensions plus ANALYTICS are only visible through the Contrail UI. In practice each type of user interacts with a SINGLE UI.

Currently most deployments use Horizon as a tenant UI and the Contrail UI as a network admin/secops UI. There are Horizon extensions for API extensions, but admin-centric API extensions plus analytics are only visible through the Contrail UI. In practice each type of user interacts with a single UI.

No. Contrail implements an IPv6 OVERLAY. No PLANS at the MOMENT to work on an IPv6 underlay.

No. Contrail implements an IPv6 overlay. No plans at the moment to work on an IPv6 underlay.

Dynamic ROUTING. CONTRAIL is built around the concept of being ABLE to FEDERATE clusters managed by potentially different admin AUTHORITIES via BGP. That is what it uses to federate its control-nodes internally and to interoperate with existing L3VPN network equipment.

Dynamic routing. Contrail is built around the concept of being able to federate clusters managed by potentially different admin authorities via BGP. That is what it uses to federate its control-nodes internally and to interoperate with existing L3VPN network equipment.

No, Contrail does not require Juniper hardware (or any specific hardware unless anticipated traffic requires the additional THROUGHPUT of a hardware SOLUTION). Mirantis and Juniper customers are presently working with several clients in production with non-Juniper hardware. The Mirantis/Juniper Reference Architecture contains instructions for configuring both Juniper and CISCO EXAMPLE hardware.

No, Contrail does not require Juniper hardware (or any specific hardware unless anticipated traffic requires the additional throughput of a hardware solution). Mirantis and Juniper customers are presently working with several clients in production with non-Juniper hardware. The Mirantis/Juniper Reference Architecture contains instructions for configuring both Juniper and Cisco example hardware.

 No. You can run Contrail without Juniper network gear. The switching gear just needs to be ABLE to pass IP PACKETS. For GATEWAY between overlay and external networks you can use the SOFTWARE gateway included in Contrail or any L3VPN-compatible appliance that supports MPLS over GRE.

 No. You can run Contrail without Juniper network gear. The switching gear just needs to be able to pass IP packets. For gateway between overlay and external networks you can use the software gateway included in Contrail or any L3VPN-compatible appliance that supports MPLS over GRE.

Yes. The CONTRAIL compute NODE components can work in case all the CONTROL-nodes became unavailable. Likewise the control nodes can continue to OPERATE in case all the config nodes become unavailable. Components are DESIGNED to operate in active-active mode.

Yes. The Contrail compute node components can work in case all the control-nodes became unavailable. Likewise the control nodes can continue to operate in case all the config nodes become unavailable. Components are designed to operate in active-active mode.

CONTRAIL is a project that IMPLEMENTS network virtualization as an overlay network compatible with EXISTING IETF standards. It is an implementation of the Neutron API that provides DISTRIBUTED ROUTING and interoperability with network appliances.

Contrail is a project that implements network virtualization as an overlay network compatible with existing IETF standards. It is an implementation of the Neutron API that provides distributed routing and interoperability with network appliances.

This depends on APPLICATION requirements. Simple APPS that can run on Linux on a KVM virtual machine can usually port to OpenStack without modifications. For higher-order applications that run on multiple servers and provide their own high AVAILABILITY using common tools like message queues and HAproxy/load balancing, it may be appropriate to architect deployments to reflect OpenStack best-practices for availability (i.e., take in mind fault domains and availability zones to eliminate likely single points of failure), but deeper changes probably won’t be needed. Only in cases where application HA is platform-dependent may deeper reengineering be needed. A GOOD introductory text for modifying applications can be found at: http://pivotal.io/platform-as-a-service/migrating-to-cloud-native-application-architectures-ebook

This depends on application requirements. Simple apps that can run on Linux on a KVM virtual machine can usually port to OpenStack without modifications. For higher-order applications that run on multiple servers and provide their own high availability using common tools like message queues and HAproxy/load balancing, it may be appropriate to architect deployments to reflect OpenStack best-practices for availability (i.e., take in mind fault domains and availability zones to eliminate likely single points of failure), but deeper changes probably won’t be needed. Only in cases where application HA is platform-dependent may deeper reengineering be needed. A good introductory text for modifying applications can be found at: http://pivotal.io/platform-as-a-service/migrating-to-cloud-native-application-architectures-ebook