Building Trust in NFV Service Lifecycle Automation
- Isn’t every vendor hyping up automation? – But how can you identify the best solution?
- How much automation do we need? – And how can we trust that the automation is working?
- Are the automation claims really possible? – Is it too good to be true?
Accanto frequently hears these concerns from communications service providers (CSPs) when discussing NFV service automation. Once CSPs actually launch NFV-based services, there is a more pressing need to address these concerns and get started with automation. A speaker from Turkcell at TM Forum’s event confirmed the challenge: “We have put a lot of effort into onboarding VNFs onto our VIM but in reality the hardest part will be lifecycle management, managing upgrades at every release, handling repairs, ensuring optimal cost per bit, etc.”
When NFV-based services are commercially available with paying customers, suddenly service providers need to be able to handle hundreds of thousands of VNF service configurations in real-time at scale. According to a recent Heavy Reading survey, 58% of service providers say that more than half of the processes that support NFV operations rely on manual scripting.
Maintaining the status quo of complex operations, manual processes and workflows is no longer an option. From a business point of view, Telecom CEOs and CFOs are more focused on how much it costs to operate networks vs. profitability of services as core revenue flattens or declines. Many of these tasks are boring and repetitive for skilled operations staff and experts who could be freed up to solve more strategic operational issues or be retrained to learn cloud software. There are strong efficiency reasons to look at software that can automate NFV service operations and end-to-end processes.
Automation is not equal: Intent vs. Traditional Service Lifecycle management
When it comes to service lifecycle management, every vendor talks about automation. One big difference is whether it’s traditional workflow-based or an intent engine. If you are automating processes based on writing scripts and coding the manual process with workflows every step of the way, then it’s workflow-based.
An intent engine abstracts away complexity by describing how a ‘service should work’ and letting the intent engine get you there. The engine breaks down the tasks, so you don’t need to program or write each process and all the possible inter-relationships and configurations. This saves time and simplifies service testing and validation by creating test scenarios, rather than writing workflow scripts. Maintenance lifecycle operations can also be automated e.g. start active/passive probes, or start analytics, hiding complexity from operations staff. (See Figure 1)
Figure 1: Intent engine capabilities vs traditional workflow capabilities
With an intent-based service lifecycle system, VNF onboarding is standardized for all VNF software components, which makes it possible to fully automate and manage services across multi-vendor VNFs. Service designers also save time by not having to write complex workflows to prescribe how heal, scale and upgrade should be handled.
A VNF service chain can be adapted for service behavior testing by calling one API resource manager. For example, you may decide to do functional testing, generate live traffic, deploy a probe, bring up VNFs, EPC, IMS etc. It’s also possible to take VNFs and service chains through a lifecycle of different states from healthy, broken, even to triple traffic pressure in order to automate the heal, scale, upgrade process. Once services are deployed, ongoing real-time analytics provide intelligence on anomalies and provide the ability to drill down and identify root causes.
Intent systems such as StratOSS™, provided by Accanto, require no manual workflow and can eliminate 30-40% of operational cost to maintain a service the duration of its lifetime, i.e. upgrade, scale, heal a service.
Trusting that automation is working
Visualisation and being able to understand what is happening is key to trusting software to automate how VNFs are started, installed and upgraded. Accanto’s StratOSS™ platform allows operations staff to see the steps the intent engine is performing; to break down the current state of the service; to view which VNFs are present and their relationships; and to understand the parameters they should have. For example, an engineer can see all the tasks that are live, as well as choose, for example, the way in which a particular upgrade gets done.
“The NFV environment needs to restore customer service itself and be service-aware, manage the customer’s service and be responsible for the service. With virtual network functions on shared resource hardware and provisioning of virtual instances and service chaining, we cannot see inside the box and need to be able to monitor the instances – up and down. We need to equip our people with tools to drill down on the service chain at the VNF level. Analytics and diagnostics are important, but should not be anonymous, we should be able to understand at the service level why something is happening so network engineering can check.”
Tier 1 Asia/Pacific national service provider
Automate at scale for production networks
While Heavy Reading’s survey also reveals over 60% of CSPs are still kicking the tires around NFV, about one-third (32%) have NFV-based services in production and are becoming serious over how to efficiently manage and operate these services. For CSPs that are already deploying NFV in live networks key challenges include managing interoperability and performance of individual VNFs, particularly upgrades involving multiple vendors.
With vEPC deployed and 5G coming, CSPs need to be able to reliably automate patch software upgrades for multiple vendors. Similarly, as part of CORD initiatives, how to manage and scale the deployment of VNFs at the edge in hundreds of data centres. Enterprise services and deploying complimentary VNFs to SD-WAN such as virtual firewalls are also a top priority. Intent-based service lifecycle management plays a key role in managing devices, spinning up CPE in the cloud, configuring virtual switches, service activation testing, etc.
While the ability to scale VNFs on-demand as more capacity is needed, is one of the greatest potential benefits of NFV, it’s also considered a top challenge to get right. Proving and trusting that NFV service lifecycle automation can work at scale comes from using and testing the software.
Accanto is happy to share more about our recent deployments and customer references, and how we can meet your requirements.
Headquartered in Finland and operating globally, Accanto provides a software platform called StratOSS™ which helps customers to simplify and automate the production of NFV services. In providing tools that unify engineering and operations, we decrease the time it takes to design, test, and simplify the production process, improving diagnostics, leading to smarter operations. www.accantosystems.com
Copyright Accanto 2018