LumenVox has communicated to all clients and partners using our legacy product architecture that we will be sunsetting this in the near future. The new architecture is built using microservice containers and incorporates our new DNN models. The new architecture is deployed using Kubernetes. In order to assist in the migration process, the following information should assist in answering any questions you may have on the new architecture and migration required.

Overview

Why is  LumenVox moving to container technology?

LumenVox is excited to upgrade to a newer, more efficient and more secure architecture. We’ve adopted container technology for the following reasons:       

• To ensure that our architecture is current and future proofed with a unified stack, and to offer simplified access to multiple product resources though unified APIs.
• To better support public cloud, private cloud, hybrid and on-premise installations.
• To enable elastic computing giving you more control and flexibility over your ROI from compute resources.
• To enable seamless collaboration between all of our speech and voice biometric products, delivering state of the art capabilities.
• To enable easier implementation for partners and their customers.
• To streamline audio acquisition across multiple telephony stacks including CCaaS, CPaaS, UCaaS, etc.
• To simplify the complexity of audio spanning that has been associated with traditional passive voice biometrics.
• To meet our small, medium and large customer requirements for this technology.

How does the migration benefit my organization? 

LumenVox’s new architecture offers users a more efficient and secure experience. Additional benefits include::

• Our new architecture offers faster models and transaction processing.
• Our new DNN models are more accurate.
• Transcription is now available.
• Text normalization, redaction, punctuation & capitalization is now available.
• More ASR languages available for ASR & transcription, including: Mandarin, Cantonese, Japanese, Korean and Russian, and more are coming soon. The full list is available here: languages available.
• gRPC is now available along with MRCP, offering higher scalability and more control.
• Multi-channel audio handling is now available.
• Dialect spelling is available for ASR/Transcription.
• Access to the latest product changes, including patches and security vulnerabilities updates.
• Enhanced transcription (augmented by grammars for keyword/phrase detection) is now available.
• Access to our new neural Text to Speech (TTS) voices is coming soon.
• Access to our SaaS environment.
• The new platform will allow you to connect into other Capacity products
• Coming features, such as Sentiment Analysis and Call Summarization.

Why is LumenVox an even better choice than ever?

This is a change we made with customers, like you, in mind. We believe this solution offers an easier and more secure experience. If this change has you considering a switch, consider how LumenVox is differentiated from competitors: 

• We offer on-premise or in cloud solutions.
• Our state--of-the-art DNN models offer superior accuracy & performance.
• Our new models are also dialect agnostic.    
• We offer grammar based ASR as well as Transcription.
• Our Enhanced Transcription option is augmented by grammars for keyword/phrase detection.
• We are very responsive to your needs and support requirements, unlike cloud giants.
• Our flexible pricing solutions allow us to offer models that are right for your business.
• Both GRPC and MRCP connectivity options are available for integration.
• We are not increasing any subscription pricing in the new architecture, so the migration is as painless as possible.

Licenses & usage

Will my software stop working when the end of support is reached?

No, your existing software will continue to work, and we will still honor any existing support agreements. Please work with your account manager if you have any questions or concerns around this. Our goal is to make migration as seamless as possible for your business.

Will my software stop working when the end of life is reached?

Our plans are to migrate all users onto our new software, which has been available for a few years now. Older, legacy software is no longer being enhanced or having security patches applied, so we feel it important to encourage our users to get the best value from our products, and this means using the latest we have available.

End of Life for the legacy products is currently planned for the end of 2024, and this is why we are undertaking a wide-ranging campaign to encourage and assist customers during this migration. Once End-of-Life has been reached, when all customers have migrated, LumenVox will decommission the licensing and other support system which will mean legacy software will stop working. 

If this timing is impactful to your business needs, please work with your LumenVox Account Manager to help us understand how we can work together towards a workable timeline. We want to ensure as smooth a transition as possible for our users, as well as providing you the best we can provide, by using our latest.

What happens if I have an existing perpetual license?

Perpetual licenses are no longer available on the new DNN architecture, however the costs will be similar to existing maintenance costs for legacy perpetual licenses. Please discuss this with your LumenVox Account Manager, who will be happy to go through the various options available.

What happens if I’ve stopped paying maintenance?

Maintenance contracts for customers offer a number of benefits. One of the most obvious and most used is the ability to gain access to newer versions of our technology. This provides you a unique opportunity to upgrade to our latest and greatest product offerings. It also provides more direct access to LumenVox Support services and assistance.

If you happen to be on the fence, please allow one of our Account Managers to help you better understand how our products have improved in recent years, and why it makes sense to migrate and restart maintenance, providing you access to all of our features, as well as all security vulnerabilities patches.

Existing perpetual license users will be provided subscription licenses, which will be competitively priced at the approximate level of maintenance costs to assist in migration. Again, please discuss your specific circumstances with your LumenVox Account Manager for details of how we can customize a package that is right for your business.

Is the new software priced in the same way as the previous versions?

We are planning to maintain parity between legacy products and the new software in a way that helps all existing customers. In our new flexible licensing, we've added licensing options to be more competitive.

Our primary goal is to make the migration process as simple and sensible from an overall business perspective for customers and  partners alike. Please feel free to reach out to your LumenVox Account Manager if you have any questions, or if you are interested in exploring some of our exciting new product options, including SaaS.

How is the new software licensed?

The overall licensing works very similar to the legacy flexible licensing system, which periodically communicates with LumenVox' licensing cloud. We have added even more flexibility, based on great customer feedback, providing customers the flexibility and choice to select different products or language combinations, using a pricing model that makes sense for your business needs. Please reach out to your LumenVox Account Manager if you have specific questions.

Existing perpetual license users will be provided subscription licenses, which will be competitively priced at the approximate level of maintenance costs to assist in migration. Again, please discuss your specific circumstances with your LumenVox Account Manager for details of how we can customize a package that is right for your business.

Similar to our legacy licensing, to consume the latest LumenVox products, the system is required to communicate with LumenVox’ cloud licensing service to synchronize product utilization. This outbound connection is made using secure HTTPS, typically once per day, and contains no user identifiable information (only license usage metrics).

External firewall configurations should allow the external connection to https://licensing.lumenvox.com via port 443. For users without direct Internet access, please contact LumenVox for details of how this can be accomplished for air-gapped and similar systems.

Technical

Why did LumenVox create a gRPC interface?

We chose gRPC as our native protocol because it was designed to support highly efficient, language-agnostic communication in distributed systems. It is therefore better suited than REST for microservice-based architectures, in which individual services. This allows clients to enjoy its inherent benefits, such as multiplexing, high scalability and secure connectivity as well as its cloud-native design.

MRCP users still have the option of connecting with this protocol, however by leveraging the underlying gRPC connectivity, it allows our products to scale more efficiently, so you get the benefits of this without even needing to use it, if you prefer.

API customers wishing more direct access and control when using our products can connect directly using gRPC, which offers more capabilities than the MRCP protocol supports.

Does this new architecture support both Windows and Linux environments?

The new architecture is designed around microservice containerization, which works best and is tested in a Linux environment. We are currently testing certain installation types in windows but please engage with the LumenVox support team should you need more information on this.

How challenging is it to migrate from existing technology?

As we have embraced a different technology than the previous traditional software there may be a learning curve for customers or  partners if you haven’t been exposed to the benefits of container technology. The important thing to note is that containerization is being embraced as technology of the future by all cloud providers and large companies especially because of its benefits such as dynamic scalability and self-healing.

The LumenVox team is here to support you through the migration process. In addition to video and installation guide resources, your LumenVox Account Manager is available to provide one-on-one support. 

How long will it take to migrate my existing system and customers?

We've gone to great lengths to simplify this process as much as possible, with updated online documentation and a new dedicated knowledge base.

Timing depends on your specific use-case, including factors like number of products (ASR, TTS, Transcription, Voice Biometrics, etc.) that you will be using as well as things like the number of languages you use, and whether the infrastructure will be in your own datacenter, a cloud hosted environment (like AWS, GCP, Azure, etc.) or if you will be using our new SaaS hosted environment.

The overall size, high-availability configuration and other complexities of your system may also be factors to consider.

We've seen typical setup times vary from a couple of hours to a couple of days. After the new system is up and running, you likely want to run some test traffic through it before applying production traffic - the amount and type of testing here is largely dependent on your needs, so timing may vary.

We are here to make this process as seamless as possible, so be sure to reach out if you have any questions.

Can I install this software on-premise?

Yes, the software can be installed on-premise, in a private, hosted, hybrid or multi-cloud environment. More information can be found here Installation guides.

Where can I install this software?

You can install the software on-premise, in a private, hosted, hybrid or multi-cloud Kubernetes environment. Currently, we have installation guides for on-premise Kubernetes, Kubeadmn, Google and Amazon (EKS) documented. This can be located here Types of setups.

Do you offer a SaaS option?

We will be launching a SaaS offering starting in Q2 2024, which will be a convenient and reliable option for clients to migrate to. In the meantime we have a sandbox environment that can be used for clients to test the new software before and during migration as needed.

If you are interested in accessing our Sandbox environment, please discuss with your LumenVox Representative, so that we can provide you information and access, as well as schedule and support your needs.

If you are interested in discussing our SaaS options then speak to your sales representative or email support@lumenvox.com.

What components must I provide?

This is highly dependent on your deployment option, but there are some prerequisites for the Kubernetes deployment like MongoDB, PostgreSQL, Redis, RabbitMQ and a persistent storage device.

The requirements will also depend on your cluster and HA (high availability) requirements.

If you are an existing Windows user migrating to this new version, please see “Does this new architecture support both Windows and Linux environments?”

For more information on software prerequisites, please see:

 https://lumenvox.capacity.com/article/256167/required-external-software

Do I have to buy expensive hardware & software to run LumenVox software?

You can make use of your existing hardware, and the solution utilizes popular and well supported, open-source software components like MongoDB and PostgreSQL instead of the more expensive proprietary enterprise database technologies. We chose to support these open-source and well-supported components to assist in managing production costs.

Kubernetes, with its enhanced monitoring, scaling, DNN performance and self-healing capabilities does consume more resources than legacy installations, but your requirements will depend on your specific use-cases. Please contact us if you would like to discuss your needs at support@lumenvox.com.

Hardware requirements can be located here https://lumenvox.capacity.com/article/355935.

How secure is the product?

LumenVox utilizes secure TLS communication between components. TLS is a cryptographic protocol and is the successor of SSL. We use encryption throughout the stack (in transit and at rest). Our PKI (Public Key Infrastructure) implementation manages digital certificates for the components and public key encryption. Overall system and individual tenant security can be managed by the customer, including customer-managed keys, which can be rotated whenever needed, and needn’t be external facing.

The only external connection required is the ‘phone-home’ licensing service requirement to the LumenVox licensing server (only product usage counts are collected). See “How is the new software licensed?” For more details of this.

The upcoming SaaS offering will also be SOC2 compliant, with full SOC2 certification in the coming months as we are working closely with our external auditors to bring the new SaaS solution to the same levels of security control as the rest of Capacity.

Do I require a load balancer?

No, the Kubernetes technology has its own form of load balancer built in. LumenVox also utilizes an advanced “Service-Mesh” mechanism to optimize traffic and security within its containerized environment.

If using MRCP then load balancers are recommended from a resiliency perspective, as these are commonly used in contact centers. Having multiple mrcp-api instances can also provide inherent high-availability and resiliency. Specific configurations are customer-specific, dependent on your own needs.

Load balancers in general fall outside of the scope of LumenVox, and more within your own IT Team’s domain, however we can help plan for your specific architectural needs if required. 

How do I integrate into the new services?

Connectivity between your existing systems and LumenVox can be achieved using MRCP (if you are already familiar with this), or using our native GRPC API.

Reporting and management interfaces currently make use of REST endpoints.

For more details of our gRPC API, please refer to our online documentation: https://developer.lumenvox.com

What development is required?

During the migration process, if you use MRCP then no real changes are required. We have an mrcp-api connector that functions very similar to our previous Media Server, and is designed to be as backward compatible as possible.

Using our new, native gRPC connectivity you would need to integrate into the new APIs.

For existing C/C++ API users we recommend that you integrate using gRPC (SDKs will also be made available soon), Python sample scripts are also available to get you going.

These are available via our GitHub repo:

https://github.com/lumenvox/speech-sample-python/tree/main/protobufs.

 Can I get assistance in installing the software?

Definitely, your system integration partner should be able to assist you as your first point of contact for technical issues.

To assist during the migration process, we are also available if needed. You can reach us at support@lumenvox.com. Note that depending on the extent and nature of your needs, a paid engagement may be required, which is why we encourage you to work with your integration partners whenever possible, but know that we are always here to help.

Additional guides and videos are also available here:

 https://lumenvox.capacity.com/

Is there additional product documentation?

Yes product, installation and support information can be found here: https://lumenvox.capacity.com/.

Additional integration information along with product information is available at: www.developer.lumenvox.com

Version-specific protocol buffer definitions and sample code are available here: https://github.com/lumenvox/speech-sample-python/tree/main/protobufs.

I’m interested in moving, what are the next steps?

We have created an internal migration calendar that our support team is using to schedule migrations in a way where we will have adequate support availability for customers and partners.

Please contact us at support@lumenvox.com to reserve your spot.

What are the minimum hardware requirements?

Specific requirements are generally use-case specific, but we have created a summary of the the latest hardware information as a guide, which can be found here:  https://lumenvox.capacity.com/article/355935

Is the new architecture multi-tenanted?

Yes, the new architecture is inherently multi-tenanted by design. From a security perspective, our design implements full data segregation between tenants, as well as other security features to safeguard each users’ data.

Whether you have multiple tenants, or just have single-user installation requirements, the same software and configuration is used. This also allows you to add tenants to your configuration later if you choose, with no additional changes needed.

Do I need a logging tool?

LumenVox supports the common containerized logging architecture, allowing customers to select whichever log consolidation or filtering application they choose, without being forced to use one product over another.

This allows customers to use their own logging platform to retrieve and analyze container logs. Customers are not restricted in which logging tools or frameworks they wish to use when working with K8s, and our software is designed to allow this level of freedom.

Cluster logging levels can be set to information, warning, error, debug, which can be applied during installation, and applied to the entire system.

Individual deployments (tenants) can override the system default logging verbosity by easily assigning the desired level of logging verbosity within their own settings, which can be updated whenever needed.

How can I monitor services?

LumenVox has out-of-the-box support for Prometheus metrics, which are available for each of the containers. This does not force users to adopt Prometheus for your monitoring needs, however the Prometheus standard of metrics monitoring is widely supported across most modern tools.

Once again, LumenVox have given you the flexibility to utilize whichever monitoring systems you are most comfortable or familiar with.

The following video may be helpful: Setup Prometheus Monitoring on Kubernetes using Helm and Prometheus Operator | Part 1 - YouTube