As 5G and the telco cloud both evolve in parallel over the next few years, vendor interoperability and other reconciliations of vendor relationships will be crucial to telcos' transformative success, according to Dimitris Mavrakis, research director at ABI Research.
In a study released earlier this year, ABI Research reported that cloud-native telco solutions will proliferate over the next five years -- tracking with the ongoing development of 5G. The probable upshot, predicts ABI, is that both of these telco tech trends will create demand for a new core network that is much more advanced than telcos have today. Telco Transformation reached out to ABI's Mavrakis to talk about both the coinciding opportunities for and obstacles to telcos in this network evolution.
Previously, in part one, Mavrakis suggested that CIOs and CTOs can best position themselves for 5G success by turning their focus away from 5G and instead applying what technology they presently have to business problems and opportunities already at hand. (See ABI's Mavrakis: Live in the Moment for 5G, Cloud Transformation.)
Now, in part two, Mavrakis delves more into the issues of interoperability for both 5G and telco cloud deployments.
Telco Transformation: Where we left off last time, you mentioned open source. What are you seeing as some of the key open standards or open-source platforms for software or hardware for telcos preparing for 5G transformation?
Dimitris Mavrakis: There are many open source projects. One of them is ONAP -- AT&T and China Mobile's Open Network Automation Platform. The first release of this was just introduced a couple weeks ago, so it's very fresh. (See: ONAP Unchained With Amsterdam Software Release.)
And then there's OpenStack, there's OpenDaylight, there are many projects. But the question is: Who integrates all of these? That is the question. So AT&T has chosen to become an integrator themselves and take ONAP, take OpenStack and take everything else, and do the work themselves, which is a lot of risk for them. Not every telco in this world can do the same. The point is that there are many open-source projects and in some cases they are competing open-source projects for the same domain -- for example, ONAP and open-source MANO for orchestration. The question is: How do telcos decide which one they go with?
TT: So how do telcos decide which open-source projects to go with?
DM: It's a very complex question that takes several factors into account, including:
- Existing software and hardware deployed in the network
- Vendor relationships
- Appetite to develop in-house
- Company politics
- Network investment plans
It's not yet clear who will invest where, but it seems that there is a critical mass of telcos looking at ONAP, but it's a complex piece of software that may scare several smaller telcos.
TT: In a release announcing its Telco Cloud Framework and Deployment Roadmaps report, ABI Research gives the example of Telefonica O2 awarding a contract for a cloud-native packet core. As both open solutions and virtualization proliferate, will we see more end-to-end cloud-native solutions?
DM: I think we will see a lot more end-to-end services. Perhaps not for the Tier-1s -- say, the AT&Ts and the China Mobiles -- but we will certainly see a lot more, because it's a risk issue. Many telcos will prefer less risk, and therefore compromise on being vendor agnostic.
TT: What are the common obstacles that you see to deploying a private or converged telco cloud? And how are they being solved?
DM: The obstacles are:
- Vendor interoperability: So if they do decide to go without a leading vendor driving everything, then it's [an issue of] interoperability with the big vendors, which means that they have to manage it themselves.
- Software licensing: Because different vendors will have different software licensing models, and telcos have to manage that.
- The business case: It may be that a truly vendor-agnostic deployment is more expensive than an end-to-end single-vendor deployment -- at least to start with. So they have to justify it to their investors.
So there are many issues. And, obviously, the multi-vendor deployment is more complex than that of a single-vendor. These are important questions to solve to answer.
These deployments are almost always private. So you won't see a telco deploying their telco cloud on, say, Amazon. I mean, there are some cases where a vendor does support some network elements on a public cloud, but I don't expect any telco will go for that. It's simply too risky.
TT: Same question, but about the non-intuitive obstacles that you have seen or expect to see; obstacles that telcos don't necessarily expect.
DM: Telcos do expect vendors to be interoperable with other vendors and with open source. And in some cases in the past this has not been the case. And this is not an easy issue to solve. It is a very complex issue. But in order for the market to progress, it is necessary to have vendor interoperability and more open systems. There are many examples where telcos commissioned a specific vendor and the vendor did not provide what the telco required and then the contract was taken back.
The technology itself is not there yet, especially with a multivendor system, but it will get there. But the most important problem is the operational transformation -- the digital transformation of the telco, which is far more complex.
TT: What other key takeaways from your research and the Roadmaps report have we not covered?
DM: There is one thing that is a risk about non-standalone access (NSA): launching mobile 5G with an EPC -- a 4G LTE packet core for mobile broadband. The risk is that telcos will rush to launch this and not focus on the more important issues that a Next-Generation Core requires. For example, telco cloud, and the operational and digital transformation that's needed to address enterprise verticals -- that's a risk. But, you know, in some cases, the industry finds ways to solve these issues sooner or later. But at the moment it's not clear how.
— Joe Stanganelli, Contributing Writer, Telco Transformation