fshy

fshy / IP and proxies

NovixLink

novixlink.com · IP and proxies · 1 record · JSON

NovixLink (诺联主机) IP and proxies

caveatemailno cardDvolume-quota

novixlink.com/store/nttispipvps · ISP-registered VPS (Chinese market) · ip-proxy/novixlink-isp-vps · JSON · compare

Free tierNo free tier. LAX-CUPN Lite CAD 7.99/month: 1 vCPU, 768 MB DDR4, 10 GB RAID10 SSD, 100 Mbps up and 100 Mbps down, 500 GB bidirectional traffic per month, 1 dedicated US dual-ISP native residential IPv4, test IP 192.220.114.2. LAS-COX Lite CAD 19.99/month: 1 vCPU, 768 MB DDR4, 10 GB RAID10 SSD, 200 Mbps up and 200 Mbps down, 1000 GB bidirectional traffic per month, 1 dedicated US dual-ISP native home-broadband IPv4, test IP 24.120.*.*.
unit none; metered on none; The Cox store page states 1000 GB per month for LAS-COX Lite. The commissioning summary for this record abbreviated that to 1 GB; the page is the authority and the record carries the page's figure.
Pricenone stated
Per-object capnot applicable
Retentionnot applicable: a rented machine and its address, not stored bytes. Terms 5.3: on non-payment the VPS, its IP and its data are deleted at expiry, with a grace period of about 48 hours that is stated not to be a guarantee of resource retention.
Egress500 GB 每月双向流量 (bidirectional traffic per month) on LAX-CUPN Lite, 1000 GB on LAS-COX Lite. The vendor publishes no overage price on either page: what happens past the allowance is not stated, so the allowance is a ceiling without a stated consequence.
free_allowance: 500 GB/month on LAX-CUPN Lite; 1000 GB/month on LAS-COX Lite
Identityemail The registration form asks for a name, an email address and a password and nothing else: no phone, no card, no identity document. Payment is Alipay, card or crypto. Terms 5.2 names Alipay and crypto as methods that cannot auto-debit and a card as the one that can; 6.2 names crypto again, as a payment that can only be refunded to site balance.
Limits
  • volume-quota 500 GB bidirectional traffic per month on LAX-CUPN Lite; 1000 GB on LAS-COX Lite, 2000 GB on LAS-COX Basic, 4000 GB on Pro, 8000 GB on Ultra source
  • per-ip-shaping 100 Mbps peak up and down on LAX-CUPN Lite; 200 Mbps peak up and down on every LAS-COX plan source
Two constraints the vocabulary cannot name. First, the ISP label is only what the vendor's own IPLense reading said at first delivery (terms 4.2), so it is a claim about one database's reading of the address rather than a property of the address, and it does not move when that database changes. Second, every LAS-COX plan was shown as 全部售罄 (all sold out) on 2026-10-24, so the one line that passed the registration check could not be bought that day.
Bindingvolume-quota 500 GB a month is about eleven times the gateway's ~45 GB/month, so the quota does not fail at a user's scale; it is named because it is the only constraint here the vocabulary can express and the one the product states. The constraint that actually decides this record is the registration class, which has no vocabulary value, and it is what the verdict's caveat is about.
Interfacessh SSH to the VPS on its static address. SOCKS5 or HTTP proxying is whatever the client runs on the machine, not a vendor endpoint: the vendor sells the address and the box, not a proxy service. Orders, renewals and IPv4 changes are handled in the customer portal.
ToolingAn SSH client plus a proxy daemon the client installs on the VPS. The vendor publishes no proxy endpoint, no API and no wrapper. That is a note about tooling and not a reason to reject the provider: the address is the product, and the machine is only how it is held.
Composes withThe Worker opens an outbound connection to the VPS's static IPv4 and the VPS makes the upstream request, so the upstream sees an address in AS2914, AS3257 or AS22773 instead of AS13335.
ReliabilityThe address is static for the life of the plan, but terms 7.3 lets the vendor switch the line, rate-limit, block ports or null-route the IP for at least an hour on a DDoS or attack, and terms 4.2 fixes the ISP label to the IPLense reading at first delivery, so a later re-registration of the range does not reopen the claim.
NotesThree candidates were on the table before the measurement, and the measurement split them. (1) The ISP label is real across the brand: refuted for two of the three ISP lines, since NTT and GTT both read DCH on both fields while the Cox line reads ISP on both. (2) The label is a database artefact the vendor picks rather than a property of the wire: confirmed, and it is the reason the verdict is a caveat and not a clean pass, because terms 4.2 makes the vendor's own IPLense reading at first delivery the standard and 4.4 excludes the professional-mode view that would show two databases disagreeing. (3) The lines differ enough that one verdict cannot cover the brand: confirmed, which is why the record is written against the brand with the Cox line named as the survivor, and a later session may reasonably split NTT and Cox into two records because their price, traffic and registration class all differ. The decisive vendor sentence, from terms 4.2: '双 ISP 住宅 IP 和双 ISP 家宽 IP 仅以服务首次交付时的 IPLense 快速模式结果和系统记录作为正式交付标准'. Read against knowledgebase 48, 双 ISP means the ASN name and the company type label both read ISP on one database, not that two networks carry the traffic, so the claim this record tests is exactly the two fields that were measured. Terms 4.3 also matters for how far the measurement travels: a delivered IP need not be in the sampled /24, only in the same ASN and the same operating company, so the pass on 24.120.4.1 is evidence about the prefix sampled and about the ASN, not proof that every address the vendor hands out in that ASN reads ISP. Each line was checked with one address, taken from the test IP the vendor itself publishes; a second and third address per line would show whether the /24 is uniform, and buying the cheapest plan and repeating RIPE plus IPLense on the delivered address would settle what the samples can only suggest. Nothing in dossier/egress/ or dossier/compute/ names this vendor, so this record corrects no other record.
Evidence rungdocumented The verdict rests on plan prices, the traffic quota and the signup fields, all read from vendor pages, so documented is the weakest rung among them. The registration class of the three sample addresses is measured with RIPE, ARIN RDAP and IP2Location, but on one address per line rather than on an address the vendor delivered a purchase that produced a delivered IP and a repeated RIPE check would move the record to measured.
Checked2026-10-25 (live); stale on 2026-11-24
Parsed numbers
  • $.limits[0].value limit:volume-quota 500.0 GB
Evidence (7)
datesourcewhat it proved
2026-10-24novixlink.com/store/nttispipvpsLAX-CUPN Lite at CAD 7.99/month: 1 vCPU, 768 MB DDR4, 10 GB RAID10 SSD, 100 Mbps up and down, 500 GB bidirectional traffic per month, one dedicated dual-ISP residential IPv4, test IP 192.220.114.2. The page names NTT America as the upstream and the product 美国NTT洛杉矶双ISP住宅IP中国优化VPS.
2026-10-24novixlink.com/store/cox-isp-vpsLAS-COX Lite at CAD 19.99/month: 1 vCPU, 768 MB, 10 GB RAID10 SSD, 200 Mbps up and down, 1000 GB bidirectional traffic per month, test IP 24.120.*.*, described as 真实家宽运营商直供 (supplied directly by the real home-broadband operator). All four LAS-COX plans show 全部售罄 (all sold out).
2026-10-24novixlink.com/legal/terms-of-servicesTerms 4.2 is the delivery standard quoted in notes; 4.3 states a delivered IP need not share the test IP's subnet, prefix, /24 or neighbours but must be in the same ASN and run by the same operating company; 5.2 names Alipay, crypto and card as payment methods; 7.3 allows line switching, rate limiting, port blocking and null routing during an attack.
2026-10-24novixlink.com/knowledgebase/48/The vendor's own definition of 双 ISP: the ASN name and the operating company's type label must both read ISP on IPLense, and the article says this does not mean two lines. It also states that geography, risk score, purity, platform results and speed are not part of the delivery standard, and that IPLense 专业模式 is for comparison only.
2026-10-24probe:192.220.114.2RIPE stat: AS2914 NTT America; ARIN RDAP: NET-192-220-0-0-1, a direct allocation; IP2Location demo: usage_type DCH and as_usage_type DCH. The NTT line fails the dual-ISP claim on the vendor's own two-field test.
2026-10-24probe:216.36.86.2RIPE stat: AS3257 GTT-BACKBONE; IP2Location demo: usage_type DCH and as_usage_type DCH. The GTT line fails the dual-ISP claim.
2026-10-24probe:24.120.4.1RIPE stat: AS22773 Cox Communications; IP2Location demo: usage_type ISP and as_usage_type ISP; PTR wsip-24-120-4-1.lv.lv.cox.net. The Cox line passes the dual-ISP claim.
More fields (2)
{
 "rung": 5,
 "rung_note": "The product is a claim about how the address is registered and nothing else, so the record sits at rung 5 and the measurement tested exactly that: ASN name and usage-type, per line, one sample each."
}
Unknown or withheld values (1)
  • $.per_object_cap.stated not-applicable