The Ciena Insights Podcast

1 00:00:09,470 --> 00:00:11,390 Hi, and welcome to the latest edition 2 00:00:11,390 --> 00:00:13,630 of the Ciena Insights podcast. 3 00:00:13,750 --> 00:00:16,440 Today, we will be talking about autonomous 4 00:00:16,440 --> 00:00:18,470 networks, getting to level five. 5 00:00:18,470 --> 00:00:21,170 My name is Karina Dahlke. I'm the head of 6 00:00:21,170 --> 00:00:23,890 marketing for Blue Planet, a division of Ciena. 7 00:00:23,890 --> 00:00:26,390 I'm joined by two Ciena leaders who 8 00:00:26,390 --> 00:00:28,290 work together frequently and are very 9 00:00:28,290 --> 00:00:30,570 knowledgeable about autonomous networks. 10 00:00:30,850 --> 00:00:35,150 Gabriele Di Piazza, VP of Products, AI, Data Science, 11 00:00:35,150 --> 00:00:37,870 and Alliance at Blue Planet, and And Erik 12 00:00:37,870 --> 00:00:40,870 McLaughlin, Senior Director of Product Line Management. 13 00:00:41,050 --> 00:00:43,010 So, let's kick it off. Gabriele, 14 00:00:43,010 --> 00:00:44,080 do you want to introduce yourself? 15 00:00:44,490 --> 00:00:45,490 Hi, everyone. 16 00:00:45,490 --> 00:00:47,930 As Karina said, my name is Gabriele. I 17 00:00:47,930 --> 00:00:50,410 head up the products portfolio here at Blue Planet. 18 00:00:50,650 --> 00:00:54,030 Obviously, autonomous network is a key, not 19 00:00:54,030 --> 00:00:56,650 just a focus of ours, but it's probably a strategic 20 00:00:56,650 --> 00:00:58,830 direction for the portfolio, given that we 21 00:00:58,830 --> 00:01:01,870 operate in the space of intelligent automation, 22 00:01:01,890 --> 00:01:03,610 telecom, telecom space. 23 00:01:04,110 --> 00:01:05,690 Great. And Erik? 24 00:01:06,030 --> 00:01:10,590 Hi, folks. I lead the Navigator Network Control 25 00:01:10,590 --> 00:01:13,590 Suite, Product Line Management, and Solutions Engineering 26 00:01:13,590 --> 00:01:16,510 here at Ciena, supporting our hardware portfolio. 27 00:01:16,950 --> 00:01:17,910 Super. 28 00:01:18,110 --> 00:01:21,990 Okay, let's dive right in. A lot of people are talking 29 00:01:21,990 --> 00:01:24,950 about autonomous networks. So if you go to any trade 30 00:01:24,950 --> 00:01:27,990 show, you'll see that theme everywhere, and it is 31 00:01:27,990 --> 00:01:30,530 definitely top of mind for many of the service provider 32 00:01:30,530 --> 00:01:32,730 executives that we are meeting with. So how about 33 00:01:32,730 --> 00:01:36,590 we start by defining what autonomous networks means? 34 00:01:36,980 --> 00:01:38,950 Gabriele, if you want to take a pass at it. 35 00:01:38,950 --> 00:01:40,650 Autonomous networks is not 36 00:01:40,650 --> 00:01:42,850 a new term for the industry. 37 00:01:42,910 --> 00:01:46,970 It's been an aspirational path on how to 38 00:01:46,970 --> 00:01:51,490 basically evolve towards this self-configuring, 39 00:01:51,670 --> 00:01:52,850 self-healing, 40 00:01:53,010 --> 00:01:58,210 self-optimizing, and ideally zero-touch operations 41 00:01:58,210 --> 00:02:02,170 networks. It's a path that is achieved across multiple 42 00:02:02,170 --> 00:02:07,730 steps. Ideally you would remove manual operations across 43 00:02:07,730 --> 00:02:13,010 individual devices, domains, and for overall services. 44 00:02:13,010 --> 00:02:17,350 I think there are some core aspects here looking at 45 00:02:17,350 --> 00:02:22,470 real-time observability expressing an intent of where you 46 00:02:22,470 --> 00:02:25,750 would like your network or service to be reconciling 47 00:02:25,810 --> 00:02:29,690 this intent and of course, perform a closed-loop 48 00:02:29,690 --> 00:02:34,050 remediation or automation and over time, remove some of the 49 00:02:34,050 --> 00:02:37,270 manual interaction in these different steps. Makes 50 00:02:37,270 --> 00:02:39,290 sense. So you're taking out a lot of those manual 51 00:02:39,290 --> 00:02:44,690 processes, automating those and injecting intelligence into 52 00:02:44,690 --> 00:02:47,530 the network. Erik, I'm going to toss this question 53 00:02:47,530 --> 00:02:51,970 over to you. A few years ago, we saw autonomous networks 54 00:02:51,970 --> 00:02:54,290 everywhere. It was all the rage. And then all of a 55 00:02:54,290 --> 00:02:56,550 sudden it went out of style. You didn't see it so much. 56 00:02:56,550 --> 00:02:58,690 You didn't hear people talking about it. You didn't see 57 00:02:58,690 --> 00:03:01,410 it at trade shows. What do you think has changed in 58 00:03:01,410 --> 00:03:04,770 the last two to three years that makes autonomy feel 59 00:03:04,770 --> 00:03:08,210 like it's really achievable now? Great question, Karina. 60 00:03:08,210 --> 00:03:12,550 And we go back a couple of years when ChatGPT entered 61 00:03:12,550 --> 00:03:15,970 everyone's lives and everyone started to interface with 62 00:03:15,970 --> 00:03:19,190 it. I think the visibility of these frontier models 63 00:03:19,190 --> 00:03:23,170 like Gemini and Copilot in everyone's day-to-day lives 64 00:03:23,170 --> 00:03:26,090 really opened up the door to people starting to really 65 00:03:26,090 --> 00:03:28,310 think of how I could leverage these large language 66 00:03:28,310 --> 00:03:32,470 models in a networking use case. That brought to the 67 00:03:32,470 --> 00:03:34,630 attention of a lot of people and a lot of focus around 68 00:03:34,630 --> 00:03:38,110 networking. Networking is a prime place to drive some 69 00:03:38,110 --> 00:03:40,650 of these large language models to drive reasoning. 70 00:03:40,650 --> 00:03:44,890 They're perfect. They require a bunch of different aspects 71 00:03:44,890 --> 00:03:47,230 of the network to make a decision, whether it's 72 00:03:47,230 --> 00:03:51,610 telemetry, instrumentation from the devices, logs, alarms. 73 00:03:51,750 --> 00:03:54,110 So what better way to have something than 74 00:03:54,110 --> 00:03:55,910 to have something that can come and help and 75 00:03:55,910 --> 00:03:58,350 reason and act like your virtual engineer? 76 00:03:58,550 --> 00:04:01,550 Exactly. It's like augmenting your staff, right? 77 00:04:01,550 --> 00:04:04,910 That's right. Great. So as you know, you both 78 00:04:04,910 --> 00:04:08,090 know, very familiar, TM Forum provides a lot of tools 79 00:04:08,090 --> 00:04:11,070 for corporations to do self-assessments on their 80 00:04:11,070 --> 00:04:13,950 level of autonomy. Gabriele, what are you seeing 81 00:04:13,950 --> 00:04:16,570 in the market today? What level of autonomy 82 00:04:16,570 --> 00:04:19,530 are we seeing when we talk to customers today? I 83 00:04:19,530 --> 00:04:21,710 think this is a very interesting question, Karina, 84 00:04:21,710 --> 00:04:26,510 because there is a wide spectrum of responses to 85 00:04:26,510 --> 00:04:30,610 this. What we start to see, which is refreshing, 86 00:04:30,750 --> 00:04:35,390 is the focus actually has increased dramatically the 87 00:04:35,390 --> 00:04:38,220 last 6 to 12 months with all this self-certification, 88 00:04:38,220 --> 00:04:41,610 self-valuation on these levels i would say that we 89 00:04:41,610 --> 00:04:45,750 always see this aspect around level two to level three 90 00:04:45,750 --> 00:04:50,090 but a lot of discussion around how do we bridge this 91 00:04:50,090 --> 00:04:53,110 gap from level three to level four but let me give 92 00:04:53,110 --> 00:04:56,270 a couple of examples just in the last month we've been 93 00:04:56,270 --> 00:05:00,870 working with some fairly large carriers and one of 94 00:05:00,870 --> 00:05:04,890 those they self-evaluating themselves to be at level 95 00:05:04,890 --> 00:05:08,230 one which was a big surprise On the other hand, I 96 00:05:08,230 --> 00:05:12,100 think this is a very humble but practical assessment 97 00:05:12,100 --> 00:05:16,210 where they basically decided that they need to get serious 98 00:05:16,210 --> 00:05:20,150 in instrumenting this change. On the other hand, 99 00:05:20,150 --> 00:05:24,290 we see some advanced carriers looking at, of course, 100 00:05:24,290 --> 00:05:26,390 you need to look at this service by service, right? 101 00:05:26,390 --> 00:05:29,390 There's not a single number of valuation for this. 102 00:05:29,470 --> 00:05:34,310 Looking at having reached level four and starting to look 103 00:05:34,310 --> 00:05:37,250 at how to push towards a level five autonomy and 104 00:05:37,250 --> 00:05:41,670 already talking about managing hundreds of plus of 105 00:05:41,670 --> 00:05:45,980 agents in their networks. So you see the spread varies, 106 00:05:45,980 --> 00:05:49,590 but I think this is the natural state of the market. 107 00:05:49,590 --> 00:05:53,370 You can also think of some of these operators having 108 00:05:53,370 --> 00:05:56,550 achieved some high level of autonomy on a specific 109 00:05:56,550 --> 00:06:00,330 service and a lower level autonomy in some other service. 110 00:06:00,330 --> 00:06:02,530 That's also very, very typical. That makes a lot of 111 00:06:02,530 --> 00:06:04,840 sense. So like you said, it's a pretty wide spectrum. 112 00:06:05,030 --> 00:06:08,290 There are going to be certain instances that it would 113 00:06:08,290 --> 00:06:11,370 be a lot easier to have some quick wins and achieve 114 00:06:11,370 --> 00:06:13,910 some autonomy. But to achieve that, like level 115 00:06:13,910 --> 00:06:16,670 four, level five, you really need to go across the 116 00:06:16,670 --> 00:06:19,550 different domains, which is really, really complex, 117 00:06:19,550 --> 00:06:21,910 given the size and the complexity of the networks. 118 00:06:21,910 --> 00:06:24,540 But I would say that if I have to characterize this, 119 00:06:24,540 --> 00:06:27,830 a majority of the discussion right now revolves 120 00:06:27,830 --> 00:06:31,330 around how to bridge level three to level four autonomy, 121 00:06:31,330 --> 00:06:35,490 which means looking at cross-domain automation, 122 00:06:35,690 --> 00:06:39,470 looking at injecting AI and agenting capabilities for 123 00:06:39,470 --> 00:06:42,770 the reasoning that Erik was discussing in order to 124 00:06:42,770 --> 00:06:48,190 understand intent and act on basically having self 125 00:06:48,190 --> 00:06:51,190 -configuration based on that specific intent. Perfect. 126 00:06:51,190 --> 00:06:53,870 Yeah. Taking that intent and turning it into action. 127 00:06:54,170 --> 00:06:57,190 Erik, looking at this, obviously the journey, 128 00:06:57,190 --> 00:06:59,910 there's a wide spectrum. The journey is not something 129 00:06:59,910 --> 00:07:02,610 that's going to happen overnight. So as service 130 00:07:02,610 --> 00:07:04,350 providers are thinking about attaining this 131 00:07:04,350 --> 00:07:07,650 autonomy, what do you think is a realistic progression 132 00:07:07,770 --> 00:07:11,650 or do you have some suggestions for quick 133 00:07:11,650 --> 00:07:13,770 wins that you've seen with your customers today? 134 00:07:13,910 --> 00:07:17,850 Definitely a pragmatic progression I think is what 135 00:07:17,850 --> 00:07:21,480 we're hearing from our customer conversations where 136 00:07:21,480 --> 00:07:24,490 they're looking to see how can we embed intelligence 137 00:07:24,490 --> 00:07:28,690 into their existing workflows and existing applications. 138 00:07:28,690 --> 00:07:31,770 So the engineers that are using applications 139 00:07:31,770 --> 00:07:35,170 can see what's happening and why. I think that then 140 00:07:35,170 --> 00:07:38,730 that starts to progress towards guided actions, 141 00:07:38,850 --> 00:07:43,430 supervised actions, and then eventually we get to the 142 00:07:43,430 --> 00:07:46,790 closed-loop operation, which then introduces another 143 00:07:46,790 --> 00:07:51,710 area of test before you touch or digital twin concepts, 144 00:07:51,710 --> 00:07:54,660 which is an important piece of the whole autonomous 145 00:07:54,730 --> 00:07:57,530 network. So what I've seen with our customers. 146 00:07:57,530 --> 00:07:59,810 It's they want to take a very pragmatic approach 147 00:07:59,810 --> 00:08:04,030 around getting to that level four, level five vision. 148 00:08:04,240 --> 00:08:06,870 And there's a lot of value of what they can get today 149 00:08:06,870 --> 00:08:09,630 with some of the capabilities that the industry is 150 00:08:09,630 --> 00:08:13,250 coming out with. So switching gears a bit, what do you 151 00:08:13,250 --> 00:08:17,290 both of you think are some of the operational pain 152 00:08:17,290 --> 00:08:21,650 points that are pushing service providers towards autonomy 153 00:08:21,700 --> 00:08:24,830 today? Like we said, people talked about it for a 154 00:08:24,830 --> 00:08:28,950 long time, and now it's becoming a reality. So what do 155 00:08:28,950 --> 00:08:32,150 you think are some of those pain points that are driving 156 00:08:32,150 --> 00:08:34,930 this change that we're seeing with our customers? 157 00:08:35,390 --> 00:08:37,830 Gabriele, I'll ask you first and then turn 158 00:08:37,830 --> 00:08:40,710 it over to Erik. I think the key pain point is 159 00:08:40,710 --> 00:08:43,670 the cost pressure, that it's unprecedented. 160 00:08:43,950 --> 00:08:47,670 We've observed this over the years but we're not 161 00:08:47,670 --> 00:08:52,090 discussing anymore some incremental two, three, four, five 162 00:08:52,090 --> 00:08:56,030 percent for example OPEX reduction we know that there's 163 00:08:56,030 --> 00:08:58,950 an imbalance between revenues and costs in the 164 00:08:58,950 --> 00:09:03,050 telecom space so we start to see the desire or the 165 00:09:03,050 --> 00:09:09,310 pressure of reducing 20, 25, 30 or more on some operational 166 00:09:09,310 --> 00:09:13,250 costs. A lot of the operational costs can be attributed 167 00:09:13,250 --> 00:09:16,010 to automation therefore Autonomous Network 168 00:09:16,010 --> 00:09:19,930 becomes a recipe on how to attain that. And I would say 169 00:09:19,930 --> 00:09:23,090 that this is not just about cost, it's also about 170 00:09:23,090 --> 00:09:27,270 agility and the ability to, as a result, being able to 171 00:09:27,270 --> 00:09:32,390 introduce new services or create that cloud-like 172 00:09:32,390 --> 00:09:36,230 approach in provisioning services in the front end, which 173 00:09:36,230 --> 00:09:38,910 is a big desire for many carriers. So these are 174 00:09:38,910 --> 00:09:41,450 some of the two key aspects, cost pressure on one side 175 00:09:41,450 --> 00:09:43,710 and the agility to introduce services on the other. 176 00:09:44,060 --> 00:09:46,770 Yeah, and if I could maybe add to Gabriele's point, 177 00:09:46,770 --> 00:09:48,970 I think the complexity of networks, they're not 178 00:09:48,970 --> 00:09:52,370 necessarily getting simpler. From a hardware CAPEX 179 00:09:52,370 --> 00:09:57,230 perspective and a lot of new optical configurations 180 00:09:57,230 --> 00:10:01,650 where people are deploying coherent plugs on routers 181 00:10:01,650 --> 00:10:04,290 and mixing and matching, right? Looking for the best 182 00:10:04,290 --> 00:10:07,030 performing plugs on the market and wanting to deploy 183 00:10:07,030 --> 00:10:09,450 on their existing router base. So that might 184 00:10:09,450 --> 00:10:12,110 introduce CAPEX savings from a hardware perspective, but 185 00:10:12,110 --> 00:10:15,190 adds some complexity on the automation side and the 186 00:10:15,190 --> 00:10:17,790 software side. So that's kind of an area where 187 00:10:17,790 --> 00:10:20,570 people are very interested in looking at automation and 188 00:10:20,570 --> 00:10:23,790 simplifying those workflows and just the scale of 189 00:10:23,790 --> 00:10:26,590 networks that are growing, with AI on the network 190 00:10:26,590 --> 00:10:29,330 and training data sets, there's a huge pressure on 191 00:10:29,330 --> 00:10:32,630 optical networks across the globe. So how do we bring 192 00:10:32,630 --> 00:10:35,730 automation, not just from a monitoring standpoint, 193 00:10:35,730 --> 00:10:37,970 but also from a commissioning point? How do I get the 194 00:10:37,970 --> 00:10:40,950 equipment faster and deployed quicker in the ground? 195 00:10:40,950 --> 00:10:43,940 From putting in the ground to lights on to optimizing 196 00:10:43,940 --> 00:10:47,030 the solution for the use case that it's delivering. 197 00:10:47,150 --> 00:10:49,810 And Erik, I like your point here because 198 00:10:49,810 --> 00:10:52,290 you started by saying that networks are not 199 00:10:52,290 --> 00:10:54,810 becoming simpler these days. Configuration, 200 00:10:54,810 --> 00:10:57,210 you mentioned data, different endpoints, 201 00:10:57,210 --> 00:11:00,770 but also the amount of services. We are discussing many 202 00:11:00,770 --> 00:11:03,890 cases where they realize that there's an overlaying 203 00:11:03,890 --> 00:11:08,050 of cloud services and B2B services. And this also 204 00:11:08,050 --> 00:11:11,550 introduces new areas of what you need to manage and 205 00:11:11,550 --> 00:11:15,350 automate. So it actually becomes a matrix of complexity, 206 00:11:15,350 --> 00:11:19,070 not just one single dimension, but it becomes almost an 207 00:11:19,070 --> 00:11:22,590 exponential complexity going forward. Yeah, exactly. 208 00:11:22,670 --> 00:11:26,930 And that complexity is not slowing down at all. I was 209 00:11:26,930 --> 00:11:30,210 at a recent industry conference and a large tier 210 00:11:30,210 --> 00:11:35,130 one provider was talking about why he was moving towards 211 00:11:35,130 --> 00:11:37,470 an autonomous network. And that was exactly the 212 00:11:37,470 --> 00:11:39,710 point that he brought up. The thousands of different 213 00:11:39,710 --> 00:11:42,610 parameters in the network, the thousands of different 214 00:11:42,610 --> 00:11:45,390 devices, hundreds of different domains. It was just 215 00:11:45,390 --> 00:11:49,690 beyond human capacity to manage it all. So that's a 216 00:11:49,690 --> 00:11:53,150 big push, right? It's just that it's got beyond us. 217 00:11:53,290 --> 00:11:55,190 There's not many greenfield networks 218 00:11:55,190 --> 00:11:57,190 out there. They got to worry about 219 00:11:57,190 --> 00:11:59,010 what's in the ground already too. 220 00:11:59,170 --> 00:12:02,470 Exactly. How to tie that all together. Super. 221 00:12:02,770 --> 00:12:06,370 Okay. So if we look at, now we understand there's 222 00:12:06,370 --> 00:12:09,210 definitely a need for autonomous networks. We're 223 00:12:09,210 --> 00:12:11,670 seeing people move in that direction. We're seeing 224 00:12:11,670 --> 00:12:15,350 real examples. How do we get there? Gabriele, I'll 225 00:12:15,350 --> 00:12:17,930 toss this one over to you. What are some of the 226 00:12:17,930 --> 00:12:20,570 foundational things that you need for an autonomous 227 00:12:20,570 --> 00:12:23,810 network? If somebody wants to start their journey, 228 00:12:23,810 --> 00:12:26,530 what is one of the first steps they should take? 229 00:12:26,680 --> 00:12:30,310 I touched a little bit at the beginning where some of 230 00:12:30,310 --> 00:12:34,950 the core aspects are really starting to abstract the 231 00:12:34,950 --> 00:12:38,470 majority of your network resources. In this way, you 232 00:12:38,470 --> 00:12:42,130 can actually treat this as individual services that you 233 00:12:42,130 --> 00:12:45,870 can then compose over time. The ability to express 234 00:12:45,870 --> 00:12:49,290 an intent to start in terms of your desired state, 235 00:12:49,290 --> 00:12:52,830 having observability at different levels that can then be 236 00:12:52,830 --> 00:12:56,770 composed into something that you want to feed into an 237 00:12:56,770 --> 00:13:00,230 engine. We haven't talked deeply yet about how AI 238 00:13:00,230 --> 00:13:04,210 plays a role here and how you can interpret what actions 239 00:13:04,210 --> 00:13:09,110 to be taken. There's the understanding of the different 240 00:13:09,110 --> 00:13:12,790 domain and how each domain has its own specific 241 00:13:12,790 --> 00:13:17,290 needs and what is programmable or not. Ideally, the 242 00:13:17,290 --> 00:13:20,990 journey here, you don't go from one to five or zero to 243 00:13:20,990 --> 00:13:25,230 five just by implementing some AI, but you need to start 244 00:13:25,230 --> 00:13:27,810 looking at each of the building blocks and exposing 245 00:13:27,810 --> 00:13:31,090 and abstract some of these building blocks. This is on 246 00:13:31,090 --> 00:13:33,880 the technology side. Then there's a cultural aspect 247 00:13:33,880 --> 00:13:37,370 where you need to apply new skills these days. You need 248 00:13:37,370 --> 00:13:41,830 to look at how analytics AI and even DevOps or net 249 00:13:41,830 --> 00:13:45,190 DevOps and cloud plays a role into this. And that's why 250 00:13:45,190 --> 00:13:48,870 we see some of the most transformational projects 251 00:13:48,870 --> 00:13:53,190 here needs to be driven top down, not just bottoms up. 252 00:13:53,190 --> 00:13:56,410 That's a really good point. So it's a cultural shift. 253 00:13:56,490 --> 00:14:00,010 Yeah, adding to Gabriele's points, a lot of the 254 00:14:00,010 --> 00:14:02,890 activities that we've been spending over the last couple of 255 00:14:02,890 --> 00:14:05,610 years is what I'd call cleaning up the network, just 256 00:14:05,610 --> 00:14:08,750 getting the right visibility of what's in the network 257 00:14:08,750 --> 00:14:12,550 at the software domain. Because deploying networks from 258 00:14:12,550 --> 00:14:15,910 whatever, 5, 10, 15 years ago, it was all about how do 259 00:14:15,910 --> 00:14:18,330 we get the network up? Is it passing traffic, passing 260 00:14:18,330 --> 00:14:21,310 bit-error test, but no one's put that same amount of 261 00:14:21,310 --> 00:14:23,890 energy on the software side. So if the software 262 00:14:23,890 --> 00:14:27,270 applications can't see the data from the network, then you 263 00:14:27,270 --> 00:14:31,010 have a partial AI insight or partial insights from your 264 00:14:31,010 --> 00:14:33,350 applications. So that presents a problem. So that's why 265 00:14:33,350 --> 00:14:36,110 we've been very focused over the last little while, 266 00:14:36,110 --> 00:14:38,610 getting our networks cleaned up through a variety of 267 00:14:38,610 --> 00:14:42,030 methods. And I think once you have that foundation and 268 00:14:42,030 --> 00:14:45,330 ensuring that telemetry is coming from the network, 269 00:14:45,330 --> 00:14:47,870 enabling those telemetry interfaces off the devices 270 00:14:47,870 --> 00:14:50,670 themselves. So you're streaming data as opposed to SNP data. 271 00:14:50,670 --> 00:14:53,330 So you can have those intelligent insights flow through. 272 00:14:53,330 --> 00:14:56,310 Both of you brought up a great point that in order 273 00:14:56,310 --> 00:14:59,430 for the network to take actions and take the appropriate 274 00:14:59,430 --> 00:15:02,190 actions, the network needs to have visibility into 275 00:15:02,190 --> 00:15:05,610 the underlying data. What is actually in the network, 276 00:15:05,610 --> 00:15:09,410 the data. So if you could comment a little bit more on 277 00:15:09,410 --> 00:15:13,270 that, because I do think that's foundational and both 278 00:15:13,270 --> 00:15:17,150 Ciena and Blue Planet have done a lot of work in that 279 00:15:17,150 --> 00:15:21,370 area to make sure that we can give our customers access 280 00:15:21,370 --> 00:15:27,150 to the right data at the right time so that agents, 281 00:15:27,150 --> 00:15:30,970 the acting with autonomy, can act on the basis of 282 00:15:30,970 --> 00:15:34,790 correct information and take the appropriate actions. I 283 00:15:34,790 --> 00:15:37,690 want to go back to one of our monikers here. We keep 284 00:15:37,690 --> 00:15:42,730 saying you cannot AI what you cannot see, which was a 285 00:15:42,730 --> 00:15:45,550 little bit of an evolution of you cannot automate what you 286 00:15:45,550 --> 00:15:49,890 cannot see. Visibility is key. We start to introduce 287 00:15:49,890 --> 00:15:52,270 and discuss the topic of AI. Erik mentioned the 288 00:15:52,270 --> 00:15:56,270 observability. It's very, very important now. So network is 289 00:15:56,270 --> 00:16:00,150 an incredible source of data. It has been forever, but 290 00:16:00,150 --> 00:16:03,810 now this data can be used with different techniques. 291 00:16:03,810 --> 00:16:06,930 And by the way, this is one of the best aspects of having 292 00:16:06,930 --> 00:16:09,850 Ciena and Blue Planet part of the same company here because 293 00:16:09,850 --> 00:16:13,110 we can go so deep in the network to the network control 294 00:16:13,110 --> 00:16:15,790 that Ciena provides through the Navigator Control 295 00:16:15,790 --> 00:16:19,370 Suite, at the same time applying some of the cross-domain 296 00:16:19,370 --> 00:16:22,650 capability or cross-platform capabilities that Blue Planet 297 00:16:22,650 --> 00:16:26,890 provides. So to me, the aspect is having clean data 298 00:16:26,890 --> 00:16:30,450 is very important. It's not enough because you need 299 00:16:30,450 --> 00:16:34,730 to apply one of the key aspects that the industry needs. 300 00:16:34,730 --> 00:16:36,770 It is domain knowledge. You need to understand the 301 00:16:36,770 --> 00:16:39,430 domain. You need to understand the constructs of this 302 00:16:39,430 --> 00:16:42,730 data. You need to understand the entity relationships. 303 00:16:42,910 --> 00:16:45,790 And only through that, you can extract the 304 00:16:45,790 --> 00:16:47,850 right insights, whether those insights are 305 00:16:47,850 --> 00:16:51,650 related to planning aspects or operational 306 00:16:51,650 --> 00:16:54,890 aspects or customer experience aspects. 307 00:16:54,890 --> 00:16:58,450 So I would say this is a key aspect. In our world, this 308 00:16:58,450 --> 00:17:02,150 happens through, obviously, deep integration in how 309 00:17:02,150 --> 00:17:05,170 we keep our teams aligned. Erik and I, we work very 310 00:17:05,170 --> 00:17:08,730 closely together. So right off the factory, you can 311 00:17:08,730 --> 00:17:12,830 say that we have our platforms being integrated. And 312 00:17:12,830 --> 00:17:15,230 it's actually refreshing to see how this is evolving 313 00:17:15,230 --> 00:17:19,610 from having been integrated API levels and data level 314 00:17:19,610 --> 00:17:23,550 and now evolving into deep agentic level as well. 315 00:17:23,550 --> 00:17:27,290 So we can actually apply advanced reasoning through 316 00:17:27,290 --> 00:17:30,170 agents that are basically interacting with each other. 317 00:17:30,250 --> 00:17:32,090 One of the points you brought up, Gabriele, is on 318 00:17:32,090 --> 00:17:34,370 domain knowledge. I think that's a huge thing of 319 00:17:34,370 --> 00:17:38,430 providing strong context to networking from a networking 320 00:17:38,430 --> 00:17:42,410 company. Who better to provide actionable insights or 321 00:17:42,410 --> 00:17:48,250 autonomous capabilities on the IP optical layer than 322 00:17:48,250 --> 00:17:51,570 Ciena and Blue Planet? That's what our customers come to 323 00:17:51,570 --> 00:17:54,950 us for our expertise and a software package that can 324 00:17:54,950 --> 00:17:57,330 deliver on that and drive those insights with a 325 00:17:57,330 --> 00:18:02,250 specific view around context, around the domains. And if 326 00:18:02,250 --> 00:18:05,230 you think about it, this combination is almost 327 00:18:05,230 --> 00:18:09,470 unmatched in the industry to have the ability to span this 328 00:18:09,470 --> 00:18:12,210 knowledge from some of the business constructs or 329 00:18:12,210 --> 00:18:14,510 the way down to the network constructs. It's very 330 00:18:14,510 --> 00:18:17,970 interesting to see. Perfect. And having that rich telemetry 331 00:18:17,970 --> 00:18:21,390 fed into the upper layers of the network and to 332 00:18:21,390 --> 00:18:24,230 be able to have that two -way communication back and 333 00:18:24,230 --> 00:18:27,170 forth, provide that context is incredibly important. 334 00:18:27,170 --> 00:18:30,440 As we wrap it up, it's been a great discussion today. 335 00:18:30,440 --> 00:18:33,130 I know we talked a little bit about the end, about 336 00:18:33,130 --> 00:18:36,950 how Ciena and Blue Planet are working together. Are 337 00:18:36,950 --> 00:18:39,830 there any final closing thoughts? Erik, I'll start 338 00:18:39,830 --> 00:18:42,790 with you that you'd like to leave our customers with 339 00:18:42,790 --> 00:18:46,830 on their journey. Maybe just to reiterate Gabriele's 340 00:18:46,830 --> 00:18:49,850 point. We're under the same roof and we're executing 341 00:18:49,850 --> 00:18:53,150 on our program plans the benefit of having stuff 342 00:18:53,150 --> 00:18:56,410 being tested prior to going to the market, conversations 343 00:18:56,410 --> 00:18:59,530 across product line management and across our 344 00:18:59,530 --> 00:19:02,750 customer base, looking at key use cases that address 345 00:19:02,750 --> 00:19:05,830 from the device to the OSS level. And we're on our way 346 00:19:05,830 --> 00:19:08,910 to having that full agentic framework from the device 347 00:19:08,910 --> 00:19:12,050 level all the way to the OSS level across the 348 00:19:12,050 --> 00:19:15,070 portfolio. Great. Gabriele, any closing thoughts? The 349 00:19:15,070 --> 00:19:19,110 other aspect is it's also very important to understand 350 00:19:19,110 --> 00:19:22,430 the type of governance that we put in place. In the 351 00:19:22,430 --> 00:19:25,190 past, we discussed about devices, then evolved into 352 00:19:25,190 --> 00:19:28,810 software platforms. Now we are injecting this AI 353 00:19:28,810 --> 00:19:33,630 domain, which means how do you secure, how do you govern, 354 00:19:33,810 --> 00:19:38,330 how you manage a lifecycle, some of these aspects. And 355 00:19:38,330 --> 00:19:42,330 this is also key for the success of this evolution to 356 00:19:42,330 --> 00:19:45,530 make basically the automation components moving from 357 00:19:45,530 --> 00:19:48,890 reactive to proactive and eventually to prescriptive. 358 00:19:48,950 --> 00:19:51,770 So you can have true simulation, gaining 359 00:19:51,770 --> 00:19:56,070 confidence and so forth to move towards a level five 360 00:19:56,070 --> 00:19:59,050 of autonomy that would remove some of the 361 00:19:59,050 --> 00:20:02,510 manual actions or any manual actions out there. 362 00:20:02,750 --> 00:20:05,740 Great. Well, thank you both. And with that, I will 363 00:20:05,740 --> 00:20:08,910 wrap it up. I want to thank both of you for sharing your 364 00:20:08,910 --> 00:20:12,810 insights and really for working daily with our customers 365 00:20:12,810 --> 00:20:15,550 to support them in their autonomous networking 366 00:20:15,550 --> 00:20:19,410 journeys, both at the OSS layer and the control layer. 367 00:20:19,410 --> 00:20:22,090 So thank you for the great conversation today. 368 00:20:22,490 --> 00:20:23,570 Thanks, Karina. 369 00:20:23,790 --> 00:20:27,790 And to our listeners, you can subscribe to our Ciena 370 00:20:27,790 --> 00:20:31,590 Insights podcast on Apple Podcasts, on the public 371 00:20:31,590 --> 00:20:35,950 radio apps for Android users, and now on Spotify. So 372 00:20:35,950 --> 00:20:39,330 you can also find all of our podcasts on our website at 373 00:20:39,330 --> 00:20:43,590 ciena.com slash podcast thanks for your time today