What’s new in the telco provider landscape in India?

In the last two articles I have explored what it takes to design and build a metro optical network and the utility fiber landscape and associated dynamics of network acquisition and the crucial role it plays in metro and long-haul networks. This article is the concluding piece and explores the new developments in the telco provider landscape that are unfolding, some of which are likely to influence dynamics of network build and enterprise service offering going forward. It will also determine the get-started and grab-market-share play in the pursuit of profit for the new entrants with fresh rethink on addressable opportunities. Undeniably, telco provider narrative remains incomplete without hyperscaler led bespoke managed fiber solutions, where the focus is shifting from metro to long haul. ?

The telco providers have come a long way in the last two decades. Some have over the years developed a large and diverse product portfolio. To support that they have a complex service lifecycle management platform that’s anything but flexible and agile. Large workforce manages the quote-to-cash workflow under the shadow of larger-than-life processes with rigid assembly line production workflow. The new entrants, aware of the telco fatigue and sluggish approach, are trying to make a mark for themselves. They have limited capital to invest, but deeper focus and greater drive to make a difference and gain market foothold. Hence, their market play and product offerings are primarily focused on four network services, (1) 1/10/100G point-to-point metro and long-haul links, (2) bespoke managed fiber solution primarily targeted at hyperscalers and, (3) as-a-service offerings that empowers customers to control network usage, and (4) dark fiber lease for mid-sized ISP/IXPs. Is it old wine in the new bottle or something meaningfully different. For it to be the latter, the key is to figure out which features to add and which to drop to create the next best network service offering. It’s not instinctive but an iterative process. The new entrants are firing efficiently on all cylinders to make it work. The necessary foundation is the underlying fiber infrastructure and automation-ready operations framework. It is being built in phases across the top-7 where most data center builds are happening, gunning for the much-awaited milestone – The-Next-1GW.

Express Route Fiber: Undisturbed Reliability, Shorter Route, Lower Latency

Telco providers have large swaths of legacy fiber infrastructure, built over the last two decades. These were built with a combination of new fiber laying, fiber swaps, dark fiber IRU and utility fiber lease. Many of the network infrastructure veterans now on the other side of 50 years have had their golden period planning and rolling out these networks. Maximizing coverage was the primary objective and that determined the fiber route length. This perspective has changed. The differentiator with respect to fiber footprint is derived from shortest path express route that does not involve multi-point add-drop that is known to introduce latency and points-of-failure. ?Capacity add-drop is mostly attributed to multi-tenant tower fiberization by MNOs, where tower connectivity is daisy chained in a ring topology. This is followed by ring capacity upgrades as traffic volumes increase. It also involves ring split and ring shortening for better traffic management. These planned and unplanned activities introduce vulnerabilities slowly and silently, that eventually add up to impact network reliability and call for frequent O&M interventions. That’s where express route fiber gains credence as it remains largely undisturbed for longer period and retains its reliability. Further, long haul network builds are aiming for fish-bone architecture, mostly associated with submarine cables, for the shortest path express route, with cities enroute connected through breakout branches. If this could be combined with utility fiber segments for – shortest-path express-route utility-fiber, it would be truly unique and a prized differentiator.

Express-Route Long-Haul Utility-Fiber from NHAI – Poised to be the Game Changer!

A much-awaited market disruption for express route fiber is dark fiber lease from NHAI along Mumbai-Delhi Expressway, offering the shortest route length of 1,380 kms, rather 1,480 kms with last mile extensions. It’s expected to be available in Q2 CY2025. Comparatively the route length from telco providers is 2,100-2,300 kms. For the first time a single-window fixed-price long-haul dark fiber lease is expected to be available. Enough fibers are already laid, reportedly 8x96 fiber-cores, to ensure sufficient on-demand availability. It will create a level playing field of established and new players and make long haul managed fiber solution a lot more attractive, commercially, and operationally.

However, it’s not a done deal yet. It’s not clear if NHAI will frontend the fiber asset monetization and manage its O&M lifecycle or appoint licensed providers through a tendering process that will manage the fiber asset monetization. Reportedly NHAI is considering three wholesale partners that can further sublease the fiber pairs to interested buyers. NHAI will determine the fixed sale price, that is expected to be of the order of INR 100,000/fiber pair/km/annum. This will avoid the long-drawn painstaking tendering process and anxious wait for the results to be declared that’s been the standard operating procedure with utility fiber acquisition.

Once available, long-haul fiber route design will have three options – (1) primary route based on single span dark fiber lease from NHAI (2) secondary route of stitched OPGW segments (3) tertiary route of shortest path underground fiber. The 4th route can have two options, in some long-haul routes like Mumbai-Chennai, two diverse OPGW routes have become feasible in Maharashtra, Telangana and Andhra Pradesh. Else a hybrid route with combination of underground, OPGW, and Oil&Gas pipeline fiber could be cobbled. ???

Raw Dark Fiber Lease for Quick Start Revenue

Express route or otherwise, how feasible is it for the ambitious new entrants to offer raw dark fiber lease, in a metro. It can be argued that it’s a get started approach to keep order books ringing till the complete network is lit up and reaches operational readiness. However, given the fiber laying costs are well known, the prorated cost-plus commercials end up being less remunerative. There is also a price standardization across cities at INR 40,000-45,000/fiber pair/km/annum, with 5% bipolar variance. Further raw dark fiber doesn’t really have the essence of end-to-end network service that is monitored and managed, though RFMS if deployed can do some parts of it. The target customers are mid-sized ISPs or IXPs looking for steal deals for dark fiber lease to interconnect with data center hosted Peering Nodes and IX Nodes. Several regional and sub-regional ISPs with FTTH footprint in Tier3 and Tier3 cities are growing steadily. They focus on underserved population residing in unstructured residential pockets with pent up demand, that the big telco providers have largely ignored. Their territorial dominance with demarcated no-entry-zones for rival ISPs is a market play of a different kind, with nearly everything happening over phone calls. Where a WhatsApp message is more sacrosanct than a signed contract. They need to peer with YouTube, Facebook, Instagram, and OTT platforms, with capacity requirements that are fast approaching 100G and multiples thereof. Such providers are progressively going for raw dark fiber lease. Hence the demand of dark fiber lease from Delhi NCR to nearby cities like Meerut, Agra, Lucknow, Jaipur, Panipath are on the rise. The same holds good for Mumbai, Chennai, and Bangalore and hopefully Kolkata. It’s interesting to note that dark fiber lease can only be offered with wholly owned greenfield underground fiber. Other network acquisition arrangements prohibit sublease of dark fiber unless it’s specifically agreed to. Hence the scope and the total addressable market of raw dark fiber lease remains limited to some extent.

Bandwidth on Demand over Optical Transmission Network

Bandwidth on demand (BoD) solution for 10G and 100G links is being marketed as an innovative solution built on the marvel of NMS platform and its autonomy to control network elements. Positioned as solution for unforeseen or once-a-month additional bandwidth requirements. The new market entrants are trying to create differentiation with no nonsense and flat hourly billing for on-demand bandwidth upgrades. It certainly is better than the 95th percentile usage-based-billing model for IPT and even MPLS, that never really had mainstream acceptance. The sales team looked at it with cautious optimism fearing existing customers could ask for it to lower monthly or quarterly billing. Scheduled or unscheduled step-up and step-down of 10G or 100G links through self-service portal with service provisioning in near real time, sounds interesting. ?The fundamental question is, are customers really interested in capacity upgrades for a few hours or a couple of days. Do they run the links with near 100% utilization, without any provision for unforeseen spillovers and end-of-the-month or festivity spikes. That’s highly unlikely. Is it not the standard practice to mark the links for upgrade if it exceeds 60 or 70% utilization for a quarter. Further, do they have a pre-approved budget for such hourly bandwidth usage. Are they not required to navigate approval process involving finance and SCM to authorize such unplanned expense. Hold on, let’s do the math, it may not be that challenging. With metro 10G link available for INR 1,600,000/annum, per-hour price is INR 183. This for sure does not make sense. With 10X increase, it’s INR 1,830/hour. One of the ways it could work more effectively is to bundle a certain number of upgrade-hours with the bandwidth capacity being contracted, much like prepaid top-up for mobile data plans. For example, a 10% add-on (INR 160,000/annum) could provide 120 hours of up-to-3X upgrade. Another variant of BoD could be contracting an aggregate bandwidth capacity, e.g. 200G, that can be flexibly provisioned between any end points withing the provider’s network footprint. Customers could be hesitant with prepaid model for aggregate bandwidth capacity upfront with staggered on-demand provisioning. This can be addressed by charging a fixed/recurring bandwidth reservation charge and the billing for individual links starts once they are provisioned. Is it possible to go for BoD service without the baseline 10G link. Providers certainly will not prefer it. There should be some assured revenue. Also, while the network capacity can be provisioned near real time, the cross-connect from MMR to customer rack would take much longer. The obvious solution is pre-provisioned cross connects. To sum up, does BoD service mean serious revenue, or its fancy bells and whistles attached to point-to-point links. In all, BoD solution has its advantages and practicalities, and has yet to become a mainstream solution. It remains to be seen how effectively the new entrants can capitalize on it.?

Long Haul Managed Fiber Solution – Feasibility for Four Diverse Routes

Hyperscalers have reasonably sufficient metro optical footprint with managed fiber solution. They are now focused on replicating the success story for long haul links. Is it practicable to have four diverse long-haul routes between Delhi-Mumbai and Mumbai-Chennai, that are currently the most sought-after long-haul routes. Hyperscalers are insisting on it, just as they did for metro optical networks. It certainly is a formidable challenge and calls for the grandmasters of network planning to fold their sleeves for the needed groundwork to establish fiber route feasibility. Involves synchronized engagement with multiple providers, public and private, to orchestrate parallelized network acquisition. At one extreme, all routes cannot be underground. The other extreme is - all the routes cannot be on utility fiber either. That many utility fiber routes are not available at present. The situation, however, is improving. More state transcos are going ahead with either tendering of dark fiber lease or signing up for PPP arrangements, allowing a licensed provider with OPGW expertise to lead the monetization that does not involve tendering. The central PSUs operating Oil&Gas pipelines are gearing up for fiber asset monetization. As described earlier the inter-state national highways being built by NHAI are nearing completion and have fiber laid in dedicated utility corridors. These put together will provide multiple utility fiber route options going forward. The percentage route length of utility fiber will vary. While some will be 80-90%, others traversing states where utility fiber is not available yet could be 60-70%. Going forward, underground routes can be progressively replaced with utility fiber routes, as they become available. A new era of never before route options are set to change the rules for long haul network build. ??

Can the Triumvirate of NOC, NaaS & BoD be a service differentiator!

The NOC by itself may not be much of a differentiator. It’s a must have building block for running network operations. The NOC along with O&M and FRT work hand in hand for network operations and customer service assurance, balancing reactive firefighting and just-in-time upgrades with proactive improvement and progressive automation. The NOC facility, adorned with NMS platforms and third-party tools, with imposing video walls showing network maps and dashboards for alarms and alerts, has remained a coveted platform to impress visiting customers and instill confidence on well managed network. All this is well known. What has changed. Two things. Focus on automation from the very beginning, to minimize human errors in inventory management, resource allocation, service provisioning and performance monitoring and analysis. Empower customers to make on-demand connectivity service provisioning, spin up a virtual network appliance like router, firewall or load balancer, and put them together with point-and-click to connect with one or more cloud platforms. This calls for integration, automation and of course investment. Further, customers now ask for performance reporting on Layer3 SLA parameters (latency, packet loss, jitter) for 10G and 100G links. In addition, they ask for frame loss and flaps in the circuits provisioned over the optical transmission network. These can be viewed in real time from the customer service portal and downloaded as reports. It’s worth taking a pause and pondering, why is the customer required to view the performance reports frequently when the network delivers exceed-SLA performance consistently. ?Is it obsessive self-assurance or performance degrade anxiety or exercising proactive supervision that all-is-well.

Network-as-a-Service (NaaS) has emerged from neo-natal care as a sunrise service offering with the promise to provide empowered experience to customers for on-demand service provisioning of 1/10/100G links. Broadly, its consumption of cloud native network infrastructure that is primarily virtual network functions (VNF) and integrated bandwidth-on-demand connectivity solution through per-hour subscription model. The customers remain impervious to the underlying hardware, software, management tools, licenses, and lifecycle services. Put simply, they enjoy driving the car without bothering about how the airbags operate, provided they haven’t been through the nightmare where it didn’t open when it had to. The point to note is, being first-time-right will play an important role in building the reliability perception. Meanwhile, a few commercial deployments have taken place. Possibly some customer testimonials are also available. It, however, is too early to comment on its success, at scale. Let’s give it a year or two.

Are Hyperscalers not going for capacity lease anymore!

Are the hyperscalers done with contracting network capacity and only going for managed fiber solution. That’s not the true picture and more so for long-haul connectivity. Long-haul managed fiber solution with stitched up utility fiber segments is still a formidable challenge, that has long lead time for staggered network acquisition, deployment, and stable operations. To plan and deliver 4 diverse routes of long-haul managed fiber solution is many times more challenging. The optimal combination would be two utility fiber routes and two underground fiber routes, given three utility fiber routes are not feasible and three underground fiber routes are not acceptable. It’s unlikely that such a massive, long haul network build be operational concurrently. It’s more likely to spread over 12-18 months. Thus, hyperscalers are likely to continue with their stockpile of contracted bandwidth capacity with multiple providers in multiples of 100G or 200G that add up to multiple terabits. It will be more evident in cities like Bangalore and Kolkata, that are not yet in the priority list for managed fiber solution. Hyperscalers also seek to maintain harmonious relationships and active business engagement with multiple providers. It allows them to gather wealth of curated information and keep a check on operational efficiency. That’s how they are so deeply well informed. It however is possible that once the long-haul managed fiber solutions are delivered and its performance stabilizes, some of the leased bandwidth capacity is consolidated. Further if a single telco provider is contracted to deliver 4 diverse routes of long-haul managed fiber solution like Mumbai-Chennai, then it’s putting all the eggs in one basket. That’s where long term telco credentials will play a major role, and new providers will find it challenging to defend their position.

Taking fiber to Data Center MMR

How challenging is it for new entrants to set up network nodes in data center MMRs and connect them with three diverse fiber routes. Telos with data center business unit or controlling influence in data center business have raised pricing for the fiber pair provisioned from the zero-manhole (ZMH) to the meet-me-room (MMR). The good old practice of leasing a duct has practically stopped, that allowed the providers to blow a 48 or 96 core fiber. This has led to considerable discontent in recent times. The new market entrants, small and midsized ISPs, and IXPs are worst affected. For a new entrant with three diverse fiber routes at least three fiber pairs need to be leased. Further, this ZMH-MMR fiber pair lease has to be done every time a managed fiber solution has to be delivered. It’s perceived to be an approach by telos to make life difficult for their competitors. At times intervention from the hyperscaler is requested to intermediate if it involves delivery of managed fiber solution for them.

Data Center Clustering – ??Making Metro Optical Networks Cost Effective!

Data Center clusters are getting denser, enhancing network proximity, shortening fiber route length and making data center interconnect more cost effective. It’s as if there is an unsaid consensus on – let’s stay close. The clustering is caused in part by interest in land parcels that are closer to electricity substations, and that offer unobstructed fiber route diversity. There is also considerable interest in land parcels adjoining hyperscaler self-build sites to address spillover demand going forward. It’s speculative for sure but worth the calculated risk in some cases. This clustering leads to a unique network localization, and the fiber route management for data center interconnect within such clusters merit further evaluation. Is it possible for network providers to proactively interconnect the data centers in a cluster and be the specialized DCI provider offering dark fiber lease to other telco providers. Can Data Center operators under IP1 registration build the last mile fiber routes and offer last mile dark fiber lease. Come to think of it otherwise, 4-5 network provider trying to connect 10-15 data centers in a cluster with 3 if not 4 diverse routes, could end up becoming unmanageably excessive fiber infrastructure. Route overlaps are inevitable and with it comes unavoidable O&M nightmares. Not all data centers are created equal. Multi-tenant enterprise data centers are less demanding than hyperscaler BTS/self-build data centers from connectivity perspective.

10G and 100G pricing – Largely Stagnant Pricing, Slower Uptake!

Annual recurring charges (ARC) for metro and long haul 10G and 100G links, provisioned by default on three diverse paths, have been largely steady over the last few years. For metro, the ARC has been hovering around INR 1, 600,000 for 10G and INR 12, 000,000 for 100G. For long-haul it’s around INR 6, 000,000 for 10G and INR 35, 000,000 for 100G. These are indicative market benchmarks, with 10% bipolar variance. Understandably, it can be optimized if the number of links are higher, or contract tenure is longer. It’s interesting that the pricing for metro, long-haul and even IPL does not seem to exhibit the usual market trend of year-on-year price erosion. It’s possibly because the volume uptake hasn’t been significant yet. 10G and 100G pricing are deal specific and gets tricky because of the situational dynamics that plays out. Every provider strives to be in the race till the end and have the first right of refusal. Existing customers wanting to ramp up do have an advantageous position over new customers. However, new customers are sought after opportunities for a new entrant to grab the deal and build market foothold for itself.? It also gives rise to an interesting possibility, particularly for short span connectivity. Can managed fiber solution be a cost-effective alternative compared to bandwidth capacity lease. Simple math can bring clarity. Consider two locations connected with 2/3 diverse routes with total route length of 50kms. Dark fiber lease @INR 45,000/fiber pair/km/annum is INR 2,250,000/annum. Another one-time price of INR 1,000,000 for pair of optical eqpt. It can provide multiple 100G links. It’s worth serious consideration and check at what route length does it tip the scale.

New Hyperscaler led Submarine Cables will alter Metro & Long-Haul landscape!

New submarine cables, MIST, IAX, IEX, Blue Raman and 2Afria Pearl, mostly have 12 or 16 fiber pairs. SMW6, an exception, has 10 fiber pairs. They are scheduled to be RFS in the next 12 months if all goes well. The design capacity is mostly between 12 to 20 Tbps per fiber pair. They are majorly hyperscaler funded and they have acquired 1, 2 or 4 fiber pairs. Integration of single threaded subsea capacity with multi-path and multi-provider terrestrial capacity at terabit scale from multiple CLS and DC will lead to new wave of metro and long-haul network build.

There is a fundamental change in approach to cable landing. The SLTEs are distributed across multiple data centers that host hyperscalers global network gateways, to terminate the fiber pair(s) they have acquired. This leads to hub-and-spoke architecture for the fronthaul connectivity, where every fronthaul link needs to be protected over two diverse routes.

While its desirous for cable landing to be hosted in a data center, it’s not happening in most cases. The exception is the NTT led MIST cable system. Other cable systems mostly land in CLS close to seashore, that house the PFE and serve as the point of fiber pair breakout. These are no longer the quaint PFE huts of the good old days. They have a well-designed MMR facility where the submarine fiber pairs are patched with OLP equipment that enables the optical signal to be carried over two diverse fiber routes to data center where the SLTE and LIM setup is hosted.

Data centers hosting SLTE for one or more fiber pairs will be the focal point of subsea capacity breakout that needs to be made accessible across multiple cloud regions and eventually to availability zone data centers. Thus, data centers are set to become interconnect hubs as multiple network providers will vie to provide multi-terabit backhaul connectivity for metro and/or long-haul connectivity.

Need for multi-terabit long haul connectivity that blends express route utility fiber from Chennai to Hyderabad, Bangalore, Pune, Delhi, and similarly from Mumbai to Pune, Hyderabad, and Delhi, will lead to new wave of network build.

Groundwork for such multi-terabit long haul connectivity has likely reached advanced stage of readiness. Much of fiber route engineering and associated network planning has been done. Execution is either underway or will start very soon.

New wave of transition to entrepreneurial ventures, tired of telco fatigue

It’s the infamous “cabinators”, the self-attested industry veterans, approaching mid 50’s, with or without the onset of the abdominal bulge, and the egotistic done all, seen all mindset. Unmistakably identifiable from disinterested approach, sluggish responsiveness, can’t-be-done stubbornness, and safeguarding self-interest first. Enjoying the good times and blissfully waiting for the sunset signoff. It slows down progressive initiatives and demotivates teams downstream. External stakeholders get exasperated and look for alternatives. Interestingly however, when some of the cabinators that have the courage and confidence to move out of their comfort zones and reinvent themselves, tighten their belt and jump the ship to take on entrepreneurial assignments, they transform themselves amazingly to drive organization building and market strategy for successful digital infrastructure provider. Make effective use of their skills, experience and deep industry-connect to build teams, and develop and execute business plans. Their entrepreneurial leadership is making a perceptible impact in the market, as seen in the new entrants. They will be a force to reckon with going forward.

Conclusion

The new trends, new differentiators, new fiber infrastructure acquisition, new submarine cables are playing a significant role in metro and long-haul network build. It will influence the bespoke managed network solution for hyperscalers and drive the much-needed growth in bandwidth uptake by enterprise customers. New services like BoD and NaaS will go through proof-of-concept trials for customers to gain confidence. Whether they become mainstream network service and bring sizable and sustainable revenue remains questionable. The agile and nimble footed new entrants will compete aggressively with legacy telco providers to gain foothold in the market. The year ahead will be action packed in telco provider landscape like never before. To end, I will quote from a LinkedIn post by Sebastian Barros – Telcos need to be like a magician’s repertoire not revealing the secrets behind each trick but by captivating the audience with the marvel of their performance itself.