> For the complete documentation index, see [llms.txt](https://docs.console.zenlayer.com/welcome/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.console.zenlayer.com/welcome/elastic-compute/overview.md). # Overview ## Introduction Zenlayer Elastic Compute (ZEC) is a computing service that runs and scales compute instances — virtual machines today, with bare-metal instances on the way — on Zenlayer's global network. Deploy compute near your users worldwide, ride a high-performance low-latency backbone, and skip the hardware. ![ZEC Architecture Overview](/files/TrqD4AM7pbNkHsBItkPN) *** ## ZEC's Three Pillars ZEC's product positioning rests on three pillars. ### Global coverage ZEC reaches further than the hyperscalers, but never so far out that we cannot deliver production-grade network and compute. The locations we offer are the locations we can stand behind. The global footprint is exposed through the **Global VPC** — a single VPC (`VPC-Default`) that your account starts with, already present in every region, with an auto-provisioned non-overlapping `/20` subnet per region. Launching an instance in a new region is one click; there are no CIDR plans, no peering objects, no cross-region routes to install. Private traffic between regions within the same VPC rides the Zenlayer backbone. For reaching outside the VPC — another ZEC VPC, another cloud, or a partner network — **Border Gateway** is the edge router that speaks BGP and installs learned prefixes into the VPC route table. ![Global Coverage](/files/zwzlQ27SWYKGLsLKoZqX) ### Built for production Every line of code, every metric, every operational choice exists to give workloads running on ZEC a stable, predictable environment — and to give you the evidence you need to trust it. ZEC is not a VPS; it is infrastructure for production applications. Concretely: * **Hardware you can match to the workload.** Instance families span AMD EPYC (2nd / 3rd / 4th gen) and Intel Xeon Scalable (2nd / 4th gen), with 1:2 / 1:4 / 1:8 vCPU-to-RAM ratios and dedicated GPU families (z3a, z4i) for GPU instances. NVMe SSD block storage backs block devices. * **Public network billing that matches how mature ops teams plan.** Pick **Data Transfer**, **Flat Rate**, or **Aggregated Burstable 95th** per EIP — with an explicit bandwidth cap and a choice of loose or tight shaping — so cost and performance are decisions you make, not surprises you discover. * **DDoS mitigation is on by default.** Every EIP is covered by Layer 3/4 DDoS protection without configuration. Per-EIP threshold rules provide a second layer at the edge, including in regions where upstream cleaning isn't yet deployed. * **Metrics that are honest about what's happening.** Overprovisioning indicators, CPU spend, disk pressure, network drops, liveness, and GPU metrics are exposed so capacity planning can be done from evidence rather than vibes. * **Addressing you can keep.** BYOIP lets you run your own IPv4 prefix on ZEC with your own origin ASN (RPKI-gated). Existing allow-lists, sender reputation, and partner integrations stay intact when workloads move in. * **Dual-stack where it matters.** Cross-region private traffic carries both IPv4 and IPv6 inside the same VPC. ![Built for Production](/files/UhZFqtyjjJ3nIaCXFP6c) ### No middle layer The product surface is intentionally small and concept-light. When something cannot be exposed as a knob, support routes the issue straight to engineering. We do not put a translation layer between you and the people who can actually fix things. You see this in the product: * **Small, deliberate concept surface.** Cross-region connectivity exposes exactly one knob (the committed bandwidth per region pair) — no Region Links, no peering objects, no routes to install. Subnets reach each other within a VPC as soon as they exist. * **Managed objects, not toolkits.** NAT Gateway, Border Gateway, and the DDoS pipeline are managed components you configure, not boxes you patch. Redundancy is internal to each. * **Predictable billing shapes.** Where rates are committed (cross-region bandwidth, Flat Rate), they're a ceiling and a floor you can plan against. Where usage is metered (Data Transfer, 95th-percentile), the meter is documented, not buried. * **Support that escalates.** Issues that can't be resolved at the product surface don't sit in a tier queue — they're routed to the engineers who built the thing. ![No Middle Layer](/files/yLeh8m8Zcrjz90dHuuWI) *** ## Why Use Zenlayer Elastic Compute * **Elastic compute capacity** — add, resize, or release instances in any covered region without re-platforming. The Global VPC already has a subnet there, so expansion is a launch action, not a network project. * **Rich bandwidth choices at competitive prices** — Data Transfer, Flat Rate, and Aggregated Burstable 95th are all first-class options per EIP, with committed rates and aggregated pools priced to match how real workloads use the network — not a single metered default. * **A full public-cloud product, not a VPS stitched together** — VPC, EIP, NAT Gateway, Border Gateway, DDoS, BYOIP, snapshots, and a global private backbone are all built in as managed components. This is public-cloud surface area held to public-cloud build standards, not a handful of hosts in colocation fronted by a self-managed private cloud. * **Performance and stability as the foundation** — current-generation AMD EPYC and Intel Xeon hosts, NVMe block storage, dedicated GPU families, default-on DDoS protection, and honest per-instance metrics. The substrate is engineered to stay out of your way so the application is the only thing you have to reason about. *** ## Learn More Before you start, the following product guides are worth a read. * [**Public-Network Bandwidth**](/welcome/elastic-compute/networking/01-overview.md) — Data Transfer, Flat Rate, and Aggregated Burstable 95th: when to use which, how they're priced, and how bandwidth caps shape traffic. * [**NAT Gateway**](/welcome/elastic-compute/networking/01-overview-3.md) — regional managed NAT for outbound egress without an EIP per instance, plus DNAT for inbound port mapping. * [**Cross-Region Network**](/welcome/elastic-compute/networking/01-overview-1.md) — the Global VPC, per-region auto-provisioned subnets, and raising cross-region bandwidth per region pair. * [**Border Gateway**](/welcome/elastic-compute/networking/01-overview-2.md) — BGP-speaking edge router for reaching other VPCs, other clouds, and on-premises networks over Zenlayer's global private connectivity infrastructure.