Tierzero sells hosted capacity that still depends on racks, transit and repair windows

Summary

• Tierzero is verifiably active as a Southern California communications provider and as the registered operator of AS11509. Current public routing views show 14 originated IPv4 prefixes, one IPv6 prefix and two observed upstream networks, Arelion and Zayo.
• The company's current data-centre offer is not evidence that it owns a data-centre building. Tierzero describes a presence at One Wilshire and a partnership with DP Data Centers at 818 West 7th Street, where the building operator, not Tierzero, controls the critical facility infrastructure.
• Published figures such as 1-100 Gbps bandwidth, 585 TB of daily transfer, 60,000 users and a 50-100 Gbps traffic band describe products or activity. They do not disclose free rack space, committed power, spare ports, server inventory, restoration time or customer-specific usable capacity.
• Redundancy has several layers. Two transit providers can reduce one routing risk, while leaving common building entrances, cross-connects, edge routers, power systems, access circuits, customer equipment or support staffing as shared failure points.
• Buyers should make portability part of capacity planning. The public terms expose dependence on underlying facility and wholesale providers, customer equipment-return duties and potentially substantial early-termination charges, while public service pages do not set out a complete export and migration procedure for voice configuration or other hosted data.

The product is simplicity; the system is not

Tierzero's current commercial identity is broader than that of a conventional server host. Its service catalogue combines business internet, Hosted PBX and SIP voice, call-centre functions, managed firewall, Wi-Fi, SD-WAN and monitoring. Its home page makes the bundle explicit: one provider and one bill for internet, voice and managed networks, aimed primarily at Los Angeles and Southern California businesses. That is hosted capacity in a practical sense. Customers hand over responsibility for communications functions that once sat in an office telephone closet or under an internal network administrator's direct control.

The transfer of responsibility does not remove physical dependencies. It rearranges them. A Hosted PBX extension still needs a running service platform, reachable IP connectivity, number routing, power at the office, a working handset or software client and a correct emergency-service address. Managed SD-WAN still needs functioning edge hardware, at least one working access circuit and a control system capable of making the right decision. Data-centre bandwidth still needs racks, power, cross-connects and routers.

Monitoring still needs people who can distinguish an alarm from the cause, obtain site access and replace something that has failed.

That distinction matters because Tierzero sells accountability as well as connectivity. Its Network as a Service page says the company treats fail-safe internet, firewall, traffic management, LAN switching, Wi-Fi and VoIP as one service, monitored around the clock. A single responsible counterparty can shorten the time lost to disputes between a carrier, a phone vendor and an IT contractor. Yet a unified commercial relationship is not the same thing as unified physical control. Tierzero can own the customer relationship while depending on a building owner, an access carrier, a transit network, a cellular operator, a satellite provider or a hardware manufacturer to restore an underlying component.

The central question is therefore not whether the service is genuinely managed. Public evidence shows that it is sold and supported as managed communications. The useful question is which parts Tierzero can repair directly, which parts it can only escalate, and how long each path can take. A buyer that understands those boundaries can value the service correctly. A buyer that treats a single invoice as proof of a single fault domain cannot.

What is verifiable about Tierzero today

Several independent signals support current operation. Tierzero's website is active, carries 2026 copyright language, publishes current sales and support contacts, and offers quotes for internet, voice and managed services. Its about page says the business began in 1997 and reports 60,000 users, 1.4 million calls a month and 585 TB of data per day. These are company-supplied activity figures rather than audited capacity disclosures, but their specificity is more useful than a vague claim of scale.

The network identity is stronger still. ARIN's registration for AS11509 marks the autonomous system active, gives the name TIERZERO-AS11509 and records its original registration on 15 September 1998. ARIN's organisation record identifies Tierzero, lists administrative, technical and abuse contacts, and shows an organisation-record update in November 2024. Registry data does not prove service quality, but it connects the company name to durable internet number resources and operational contact roles.

Observed routing adds live behaviour to the registration. bgp.tools reported 14 originated IPv4 prefixes and one IPv6 prefix for AS11509, along with Arelion's AS1299 and Zayo's AS6461 as upstreams. RIPEstat's announced-prefix view independently returned the same total of 15 announced prefixes when checked for this report. Cloudflare Radar also maintains a current routing view for the network. These are meaningful signs that AS11509 is not merely a dormant registry entry.

An IPinfo view of AS11509 showed 118,528 IPv4 addresses across the listed netblocks, the same two upstreams, and responses from a sample of addresses measured from Los Angeles. That does not mean Tierzero has 118,528 servers, customers or immediately assignable addresses. A routed address can serve infrastructure, a customer, a virtual service or nothing active at all. Some address space can also be delegated or described under another organisation. The safe conclusion is that Tierzero originates a substantial address footprint and that parts of it respond on the network, not that every address represents saleable capacity.

There is also a regulatory history consistent with a communications provider. A 2017 California Public Utilities Commission agenda records Tierzero's application for authority to operate as a non-dominant interexchange carrier. An older Federal Communications Commission order lists Tierzero with an FCC registration number. Those records are historical and should not be read as a current performance endorsement. They do, however, help separate this Tierzero from similarly named technology businesses and confirm that telecommunications service is not a recent rebranding exercise.

Together, the website, registry and routing evidence support an operating-status conclusion: Tierzero is an active US communications and managed-network provider with a live autonomous system and a clear Southern California sales focus. They do not establish ownership of every cable or facility used in delivery, nor do they quantify the spare capacity available to a new customer on a particular day.

Two downtown addresses, two operating boundaries

Tierzero's physical footprint deserves unusually careful reading because its own data-centre page names two different downtown Los Angeles locations in close succession. The Data Center IP Bandwidth page says Tierzero has a presence at One Wilshire and offers 1-100 Gbps bandwidth, native IPv4 and IPv6, BGP support, redundant paths and direct cloud peering. Later on the same page, it describes “DP Data Centers at 818 W 7th Street” as a partner service for colocation, connectivity, IP transit and office leasing.

One Wilshire is at 624 South Grand Avenue, not 818 West 7th Street. CoreSite's LA1 technical specification identifies LA1 as One Wilshire at 624 South Grand and describes 173,000 square feet of colocation space, diverse points of entry, N+1 generators, controlled access and round-the-clock field support. Tierzero's PeeringDB record lists AS11509 at CoreSite LA1 One Wilshire. The record does not list a public exchange connection, and its network-level update date is July 2022, so it is evidence of a declared facility presence rather than a full current topology.

DP Data Centers describes a different facility at 818 West 7th Street. Its public figures include a 372,194-square-foot building, 66,000 square feet of available data-centre floor, a 5 MW dedicated substation, 2.5 MW of generator capacity with expansion potential, 1.5 MW of current critical-power availability, another 1.5 MW described for future expansion, N+1 cooling and round-the-clock security. DP Data Centers' company page says it is owned and operated by Downtown Properties, which controls all critical infrastructure at the building.

This ownership statement draws a clean line. Tierzero may supply IP bandwidth, customer service and a route into the building's connectivity market. It may bundle cabinets and power commercially through the partnership. Public evidence does not show that Tierzero owns the 818 West 7th building, its substation, generators, cooling plant or security operation. When utility power, a generator, cooling equipment or building access fails there, the facility operator holds the primary physical controls even if Tierzero remains the customer's first call.

The One Wilshire boundary is similar. CoreSite publishes the facility specification and operates LA1. Tierzero's presence there can place its routers near carriers and customers, but a facility presence does not reveal whether Tierzero occupies one cabinet or a larger suite, whether its edge routers sit in separate power zones, or whether its Arelion and Zayo cross-connects use diverse building entrances. Those details determine whether two logical paths survive one physical incident.

The two addresses may improve resilience if Tierzero uses them as separate active sites with independent equipment, power, transport and service state. Public pages do not state that architecture. The DP relationship might instead serve colocation customers while One Wilshire hosts network interconnection. That could be operationally sensible, but it is different from an active-active hosting platform replicated across two buildings. A buyer should not infer multi-site compute or voice continuity merely because two facility names appear on the website.

Installed capacity is not usable capacity

Infrastructure sellers often publish the largest number available somewhere in the delivery chain. Buyers need the smaller number that remains allocable after contracts, power, ports, failover reserve and operational policy are applied.

Tierzero advertises data-centre bandwidth from 1 Gbps to 100 Gbps. That range describes a service size or port class. It does not disclose the aggregate capacity of AS11509, the committed load on its upstreams, the amount reserved for failover, or the time required to provision a new 100 Gbps circuit. A 100 Gbps interface can exist without 100 Gbps of unconstrained end-to-end throughput. The cross-connect, router port, optics, transit commitment, remote destination and traffic policy all have to support it.

The company's 585 TB daily-transfer figure is similarly informative but limited. If “TB” is read as decimal terabytes and spread evenly over 24 hours, it corresponds to roughly 54 Gbps of average data movement. If the company intended binary tebibytes, the average would be closer to 60 Gbps. Real traffic is not even, and the public page does not define whether the figure is ingress plus egress, customer traffic only, voice data included, or measured over what period. It is therefore an activity indicator, not a capacity ceiling.

It happens to sit within PeeringDB's self-reported 50-100 Gbps traffic band, but those two public values may derive from the same organisation and are not independent confirmation.

The same discipline applies to DP Data Centers' facility numbers. A 5 MW substation is not 5 MW of customer-ready IT load. Generator capacity, cooling capacity, distribution losses, redundancy policy and existing commitments reduce what can be delivered. DP's 1.5 MW “current critical power availability” is more directly useful, but it belongs to the facility as a whole. Tierzero's page offers full and half cabinets at 1.5 kW to 10 kW per cabinet; it does not say how many are under Tierzero's control, how many are vacant, or whether high-density power is available in the required room and redundancy configuration.

Nor does a rack create compute by itself. The public Tierzero catalogue does not list dedicated-server models, virtual-machine sizes, storage tiers, GPU inventory or replacement-part stock. That absence is not a defect in a company focused on connectivity and managed communications. It does mean the evidence supports colocation and hosted communications claims more strongly than a broad claim that Tierzero operates an on-demand compute cloud. Customers bringing their own servers remain responsible for hardware lifecycle unless a contract explicitly transfers that duty.

Customers using a third-party platform over Tierzero connectivity inherit that platform's capacity and recovery limits as well as Tierzero's network limits.

Usable capacity is always customer-specific. For an internet customer it can mean committed information rate at the service address during peak hours, with the backup path included. For a Hosted PBX customer it can mean simultaneous call capacity, number availability, feature licences and the ability to survive a site failure. For a colocation customer it means powered rack units, cross-connect lead time, remote-hands availability and spare components. A useful quote should translate headline scale into those concrete quantities.

AS11509 shows transit diversity, not complete diversity

AS11509 gives customers a rare public window into one part of Tierzero's delivery system. An autonomous system can choose how to announce its prefixes, which upstream networks to use and how to steer traffic. The current routing views consistently identify two upstreams: Arelion and Zayo. Losing one upstream should, if the network and contracts are configured correctly, leave a path through the other.

That is real redundancy, but it is bounded. A BGP table cannot show whether both carrier circuits enter the building through the same conduit, land in the same meet-me room, terminate on the same Tierzero router, share a line card or depend on the same optical transport outside the building. It also cannot show whether both providers have enough spare capacity to absorb the other's peak load. The public path is evidence of routing relationships, not proof of physical route separation.

The prefix count also needs context. bgp.tools describes 14 IPv4 and one IPv6 originated prefix, equivalent to 463 IPv4 /24 blocks and 4,096 IPv6 /48 blocks by address quantity. PeeringDB lists recommended prefix limits of ten for IPv4 and five for IPv6, while RIPEstat returned 15 current announcements. Those values are not necessarily contradictory: one service may count aggregate routes, another may include recommended filter limits or more-specific announcements, and records can age at different rates. They do show why a number without its measurement method should not become a capacity claim.

Route-origin security remains a watchpoint. RIPEstat's route-origin validation endpoint returned an “unknown” state with no validating route authorisations for six large sampled AS11509 announcements when checked for this report, including 208.179.0.0/16, 64.239.128.0/18 and 2604:4bc0::/36. “Unknown” is not “invalid”: it generally means no matching cryptographic authorisation was found for the tested route and origin. It does not establish that Tierzero rejects invalid routes received from others, because publishing authorisations for one's own prefixes and filtering received routes are separate practices.

A current route-security statement from Tierzero would settle both questions more clearly than inference from public validators.

Tierzero also advertises direct peering with Google Cloud, Amazon Web Services and Microsoft Azure on its data-centre page. Public BGP summaries identify Arelion and Zayo but do not expose those cloud adjacencies, and the PeeringDB record shows no public exchange connection. That does not disprove private interconnection. Private virtual circuits, a partner's fabric or routes hidden from public collectors may not appear as direct AS neighbours.

Buyers who need cloud-path diversity should ask which facility, exchange or interconnection product carries each path; whether the cloud traffic avoids public transit; and whether the connection terminates on infrastructure separate from the ordinary internet edge.

The rack is only one part of the failure domain

The physical system begins before a server reaches a cabinet. Equipment has to be selected, purchased, delivered through a loading dock, received, inventoried, mounted, cabled, powered, configured and tested. DP Data Centers says deliveries can be made around the clock and Tierzero advertises remote-hands support through the partnership. Those features reduce travel time. They do not disclose a guaranteed response time, a stock of customer-compatible components or authority to open every customer's chassis.

For colocation customers, spare-parts policy is often more important than the building's nominal uptime. A failed power supply can be replaced quickly if a compatible unit is on-site and labelled. It can take much longer if a customer must approve a purchase, a courier must cross Los Angeles, or the replacement requires a vendor contract. Disk, fan, optic, switch and firewall failures create the same distinction. A facility can remain fully powered while an application remains unavailable because the final replaceable component is missing.

Tierzero's terms of service acknowledge that installation is not entirely within its control. An expedite fee buys priority rather than a guaranteed date. The terms also allow Tierzero to choose and change delivery technology, provided functionality remains substantially unaffected. Those provisions are common in communications contracts because local access, permits, building access and equipment supply can sit outside the seller's direct control. They also make delivery lead time a capacity variable. Service that can eventually be delivered is not the same as service that can be installed before a lease expiry, product launch or disaster-recovery deadline.

Repair windows should be divided by ownership. Tierzero can work on its routing policy and equipment. CoreSite or DP Data Centers controls facility systems and physical access in its building. Arelion and Zayo control their wider networks. An access carrier controls the last-mile circuit where Tierzero resells or manages it. The customer controls, or at least shares control of, on-premises power, wiring and devices. Effective incident management joins those clocks into one response, but the slowest necessary party still determines restoration.

Planned work deserves the same scrutiny as faults. Router software, optical changes, generator tests, UPS maintenance, cross-connect moves and customer edge upgrades may each require a maintenance window. Public Tierzero pages do not publish a standard notice period, maximum planned interruption or blackout calendar. Contract documents supplied with a quote may contain those details. If they do not, a customer should negotiate them, especially where hosted voice and internet share the same access circuit.

Hosted voice turns office dependencies inside out

Tierzero's business voice page offers Hosted PBX, SIP trunks, a call-centre platform, virtual fax and replacement of traditional analogue lines. Hosting removes the need for a customer to maintain a full PBX at each office, and it can make extensions easier to move. It also changes what a site outage means.

With an office PBX, internal calling may continue during an external network failure even if outside calls do not. With Hosted PBX, reachability to the host becomes central. A phone needs local power, usually through a powered Ethernet switch, plus a working local network and broadband path. If the primary circuit fails and the backup takes over, call continuity depends on correct session handling, quality-of-service policy, firewall state and sufficient backup bandwidth. The fact that email works after failover does not prove that live calls, emergency calls or call-centre queues behave correctly.

Tierzero is unusually direct about one consequence. Its legal provisions say emergency calling may fail when broadband or electrical power is lost, when equipment is moved away from the registered location, or when location records are not updated. The page recommends retaining a traditional telephone line for disruption scenarios. That warning makes clear that hosted voice capacity includes location data and local power, not only central software.

The 1.4 million monthly-call figure suggests a service with meaningful activity. It does not reveal busy-hour simultaneous sessions, geographic distribution, platform architecture, call-record retention or how many sessions can fail over to another site. Those are the voice equivalents of free rack space and spare power. A customer with 200 extensions and 20 ordinary concurrent calls has a different risk from a contact centre that can generate 150 simultaneous calls after a regional incident.

SIP trunking can preserve more customer control because the customer keeps an existing phone system and connects it to Tierzero's voice network. Tierzero's SIP trunk page supports that service model. Yet a customer-owned PBX introduces its own hardware, software and support boundary. Portability improves only if the customer has configuration access, documentation, number-porting rights and a tested alternate carrier route.

The most revealing recovery exercise is not a sales demonstration. It is a controlled test in which the primary broadband circuit is disconnected during a live but low-risk call period, emergency-location records are verified, incoming numbers are tested from outside networks, and the organisation records what survived. That exercise measures the combined service, including customer equipment and Tierzero's response, rather than one component's advertised availability.

Backup internet can still share the wrong things

Tierzero offers several access technologies: dedicated fibre, broadband fibre, cable, fixed wireless, 5G, LTE and Starlink. Its business internet page presents them as options that can be combined for automatic failover. This range can create strong diversity, but only when the secondary path avoids the primary path's physical and commercial dependencies.

A second wired circuit from a different carrier may still share the same building entrance or street trench. A cellular backup avoids the local cable but depends on tower coverage, radio conditions, the mobile operator's core, the customer's router and local electricity. Tierzero's 5G service page says a custom-programmed router monitors the primary router and takes over when it fails; it also notes data-plan limits and charges during extended outages. That is a practical backup design, but “automatic” does not mean “unlimited” or “independent of power.”

Satellite access can avoid local terrestrial routes, yet the dish, its view of the sky, local cabling and electrical supply become new dependencies. Fixed wireless can avoid a trench while sharing a rooftop, mast or regional aggregation point. Dedicated fibre offers a defined committed service, but Tierzero's dedicated-fibre page does not publicly map route diversity for each building.

Tierzero's Traffic Manager page says customers can bring their own bandwidth and combine locations, with pricing from $75 to $350 a month depending on breadth and number of sites. That option is strategically useful because it allows the management layer and the access-provider layer to be separated. A customer can retain a circuit from another provider as a hedge against a Tierzero access fault. The trade-off is a wider support boundary: Tierzero can monitor and switch traffic, but it cannot repair another carrier's network.

The right redundancy question is therefore specific: which single incident should the design survive? For a fibre cut, an LTE path may be enough. For an office power failure, both routers need independent battery runtime or staff need a remote-work plan. For a Tierzero edge problem, backup traffic may need a route that does not traverse AS11509. For a downtown facility incident, hosted services need state in another building. “Multiple carriers” is the beginning of the design, not its acceptance test.

Support labour is part of saleable capacity

Tierzero repeatedly states that US-based support is available 24 hours a day. Its support page publishes separate routes for billing, voice and general technical help. Continuous contact coverage is valuable, especially for smaller customers that do not operate their own network centre.

But availability of a phone number is not the same as available engineering capacity. Public pages do not state first-response targets by severity, time to engage a network engineer, field-dispatch coverage, escalation to facility staff, or restoration targets. They do not say how many simultaneous major incidents the team can sustain. During an isolated device failure, one capable technician may be enough. During a regional carrier outage, hundreds of customers can call at once while the same staff also coordinate the underlying repair.

This is why support metrics belong beside bandwidth in a hosted-capacity assessment. The important quantities include median and worst-case response, percentage of incidents resolved without third-party escalation, remote-hands arrival time, parts availability, and the time between a carrier notice and a customer update. A provider can have spare network capacity and still suffer operational congestion in its support queue.

Tierzero's commercial pitch of one support team may reduce organisational friction. Its public privacy statement also says customer information can be used for operations, support, network provision, maintenance, billing and collection, and that vendors or business partners may be involved in providing services. The privacy statement requires those outside parties to maintain confidentiality at least at Tierzero's stated level. That gives customers a broad description of third-party involvement, but not a list of every service location or subcontracted support function.

For sensitive customers, the practical questions are whether support personnel can access call records or device configurations, where logs and recordings are stored, how access is authorised, and what happens to retained data after termination. A claim of US-based support speaks to the people answering requests. It does not, by itself, establish that all service data stays in one US jurisdiction or that every underlying platform is operated solely by Tierzero.

Contracts expose dependencies that marketing compresses

Tierzero's terms are unusually useful because they describe the economic chain under the single invoice. The company reserves the right to pass through price increases imposed by providers of underlying facilities and by wholesale long-distance providers. It can otherwise change rates or terms with at least 30 days' written notice, with a cancellation right for qualifying material changes. The wording confirms that at least some service inputs are purchased from other providers and that their costs can reach the customer.

The same terms say an expedited installation date is not guaranteed and that installation is not entirely within Tierzero's control. Customers must obtain necessary building access and permits. They must arrange disconnection and payment with existing carriers. Upon termination, customers must allow Tierzero to remove its equipment or return it promptly. These are not obscure legal details; they identify the actual work required at the beginning and end of service.

Exit cost is material. If a customer terminates for a reason other than Tierzero's uncured material breach, the published terms call for an early-termination fee equal to 100% of the remaining agreement balance. A customer can negotiate a different order form, and applicable law or a specific service schedule may alter the result. Still, the public baseline means migration planning should begin before a long term is signed, not after performance deteriorates.

The privacy statement further says the service is provided “as is,” disclaims uninterrupted or universally available service, and caps cumulative liability in the stated circumstances at fees paid during the preceding 12 months. The service pages separately advertise 99.9% uptime for business internet and stronger language for fail-safe products. Those statements can coexist if an individual service schedule defines credits and exclusions. A buyer should obtain the actual service-level document for each component and read the measurement point, exclusions, claim procedure and maximum credit.

A percentage without those details is not a recovery plan.

Billing failure is itself an infrastructure risk. A disputed invoice, expired payment method or mistaken suspension can make a technically healthy service unavailable. The one-bill model reduces the number of invoices but concentrates the consequence of an account problem. Customers should require named billing escalation, notice before suspension, and separation between a disputed amount and undisputed critical-service charges where possible.

Portability is the reserve capacity a customer owns

Providers often describe capacity as what they can add. For the customer, resilience also depends on what can be moved away.

Internet portability begins with addresses and routing. A customer using Tierzero-assigned addresses will normally need to renumber when moving to another provider. A customer with portable address space and its own autonomous system can preserve more control, but must manage route authorisations, routing records and alternate transit. Most small businesses will reasonably choose provider-assigned space; they should still know which systems contain hard-coded addresses and how long DNS changes take.

Voice portability includes telephone numbers, emergency-location records, auto-attendant logic, call queues, greetings, voicemail, call-detail records, recordings, fax content, user lists and handset configuration. Public Tierzero pages do not describe a complete export format or migration sequence for those materials. Number porting is especially time-sensitive because a number can be legally portable yet operationally difficult to move without correct account data, letters of authority and coordination between carriers.

Managed-network portability includes firewall rules, Wi-Fi settings, virtual networks, traffic policies, device inventory, credentials and monitoring history. If Tierzero owns the edge hardware, the terms require its return. The customer needs a replacement device and usable configuration before disconnection. If the customer owns the hardware, it still needs administrative access and current backups. “Bring your own bandwidth” can reduce access lock-in, but it does not automatically make the management configuration transferable.

Colocation portability is the most physical. Servers must be backed up, shut down, uncabled, removed under facility rules, transported, re-racked and reconnected. Cross-connects and new transit must be ordered before the move. A high-density cabinet may not have equivalent power at the destination. Remote hands can assist, but chain of custody and maintenance windows matter. If the old contract has a large remaining balance, the customer may pay for parallel sites during the migration.

The reserve that makes all of this possible consists of current documentation, independent backups, spare addresses or low DNS time-to-live values, an alternate communications method, and enough overlap budget to run old and new service together. That reserve belongs to the customer. It is the most dependable hedge against any provider's contract, capacity or support limits.

Seven failure paths worth modelling

The public evidence supports seven concrete scenarios. None is a prediction that Tierzero will fail. Each is a way to test whether a proposed design converts supplier claims into customer resilience.

1. A rack loses power while the building remains up

A single cabinet distribution unit, branch circuit or power supply can fail inside a healthy facility. The response depends on dual-corded equipment, A/B feeds, device power design and spare parts. DP Data Centers' redundant facility systems do not prove that every customer device uses two independent feeds. The contract should state the delivered power topology and who may replace a failed unit.

2. One downtown facility becomes inaccessible

Fire-safety activity, utility work, security restrictions or a local emergency can block physical access even when some systems continue running. Tierzero's evidence at One Wilshire and partnership at 818 West 7th provide potential geographic options within downtown Los Angeles, but public materials do not show automatic service replication between them. Customers needing site survival should identify the second active location and test it.

3. An upstream route disappears

The observed Arelion and Zayo relationships offer a second path at the AS level. Recovery still depends on correct route announcements, filters, adequate remaining capacity and physical separation. Monitoring should verify reachability from outside networks, not only that a local BGP session remains established.

4. The access circuit survives but customer edge hardware fails

A failed firewall, SD-WAN appliance, switch or power adapter can disconnect an office from two healthy carriers. Tierzero's managed service can improve detection, while restoration depends on remote access or replacement stock. The customer should know the on-site spare strategy and field-dispatch time.

5. A hosted voice platform is reachable but business calls still fail

Number routing, registration, emergency-location records, queue logic or a software release can fail independently of internet reachability. Synthetic inbound and outbound calls provide a better signal than ping. A second carrier path helps only if the voice service itself has another working service instance or a forwarding plan.

6. Support becomes saturated during a regional incident

A large fibre cut, power event or carrier problem can generate many simultaneous requests. Automated monitoring may detect the fault, but customers still need prioritisation and updates. Severity definitions, named escalation and a pre-agreed communication channel reduce the burden on the first support queue.

7. Commercial separation becomes urgent

A merger, office closure, billing dispute, service change or repeated outage can force migration before the contract ends. The technical move may be possible while number ports, equipment returns and early-termination charges delay it. A written exit schedule and periodic export test make this failure path manageable.

What a capacity claim should contain

Tierzero's public footprint is credible enough to justify detailed questions. The answers should be attached to the exact service and address, because company-wide averages cannot describe a particular customer circuit or hosted instance.

For data-centre bandwidth, a useful disclosure identifies the serving facility, Tierzero router pair, port size, committed rate, burst policy, traffic measurement method, current utilisation range, failover headroom, transit providers, physical cross-connect routes and route-security practice. If direct cloud connectivity is included, it should identify the cloud, facility, product type, bandwidth and whether it shares the public edge.

For colocation, it identifies the contracting facility operator, room and cabinet class, available rack units, A/B power arrangement, deliverable kilowatts, cooling limits, cross-connect lead time, access rights, remote-hands response, approved delivery process and spare-parts responsibility. Facility-wide megawatts are background context; the relevant figure is power reserved and usable at the contracted cabinet.

For Hosted PBX, it identifies simultaneous-call entitlement, codec assumptions, platform locations, carrier diversity, backup routing, emergency-service handling, maintenance notice, call-record retention, recording export, number-port procedure and service restoration target. For managed networks, it adds device ownership, spare-stock location, configuration backup, administrative access, monitoring retention and field replacement time.

For every product, the buyer should distinguish a design target from observed operation. A diagram can show two carriers while both still share an entrance. A failover feature can exist without recent proof that it works. A cabinet can be offered while waiting for power. The strongest evidence combines contract language, a current physical or logical design, utilisation data and a witnessed recovery exercise.

The evidence grade is medium, and that is useful

Tierzero is easier to verify than many privately held regional communications providers. AS11509 is active. The company originates a material address footprint. Two upstream relationships are visible. A declared One Wilshire presence aligns with a major interconnection site. The DP Data Centers partnership names a real second building whose operator publishes unusually specific facility-power and cooling figures. Tierzero also publishes detailed service and contract pages rather than relying solely on broad slogans.

The remaining gaps are concentrated exactly where a capacity buyer needs precision. Public information does not show Tierzero's rack count, power entitlement, router inventory, physical transit routes, utilisation by edge, spare hardware, support staffing, multi-site voice architecture, recovery-time results or full data-export process. Some public claims, including direct cloud peering and “never lose connectivity” language, require customer-specific evidence before they can support a critical design.

That produces a medium network-evidence grade rather than a weak one. The identity and live routing layer are well supported. The facility partnership is identifiable. The unresolved questions concern how much of the underlying system is reserved, independent and recoverable for one customer.

Tierzero's simple proposition can still be valuable. A Southern California business may reasonably prefer one accountable team to a collection of unrelated carriers and equipment vendors. The service becomes stronger when the simplicity is backed by explicit boundaries: which building, whose power, which two paths, what spare device, what response time, what maintenance notice and what export route. Hosted capacity is not immaterial. It is physical infrastructure, contractual rights and human repair work presented through a cleaner interface.

The buyer's job is to make sure the clean interface does not hide the only failure that matters.