{"id":1832,"date":"2026-06-28T07:28:44","date_gmt":"2026-06-28T07:28:44","guid":{"rendered":"https:\/\/rdp.in\/gpu-mart\/?post_type=product&#038;p=1832"},"modified":"2026-07-05T09:21:18","modified_gmt":"2026-07-05T09:21:18","slug":"draco-64-port-800g-infiniband-switch","status":"publish","type":"product","link":"https:\/\/rdp.in\/gpu-mart\/product\/draco-64-port-800g-infiniband-switch\/","title":{"rendered":"DRACO 64-Port 800G InfiniBand Switch"},"content":{"rendered":"<p>The DRACO 64-Port 800G InfiniBand Switch is the high-bandwidth interconnect that turns a pile of GPU servers into a cluster. It provides 64\u00d7 800G OSFP (XDR) at 102 Tb\/s of non-blocking switching over InfiniBand XDR (Quantum-X800 class), with &lt;600 ns port-to-port \u2014 the low-latency, lossless fabric that lets distributed AI training scale across nodes without the network becoming the bottleneck. On-premises, in INR, on a GST invoice.<\/p>\n<p>Engineered as the fabric layer of an RDP AI cluster, it carries the collective (all-reduce) traffic that dominates multi-node training, and is delivered configured, cabled and validated alongside the compute and storage it connects.<\/p>\n<h3>Key highlights<\/h3>\n<ul>\n<li><strong>64\u00d7 800G OSFP (XDR) \u00b7 102 Tb\/s<\/strong> \u2014 non-blocking switching to wire a GPU cluster at full bandwidth.<\/li>\n<li><strong>InfiniBand XDR (Quantum-X800 class)<\/strong> \u2014 RDMA, lossless, in-network compute (SHARP) for fast collectives.<\/li>\n<li><strong>&lt;600 ns port-to-port<\/strong> \u2014 low latency so distributed training scales with near-linear efficiency.<\/li>\n<li><strong>800G \/ 400G per port<\/strong> \u2014 flexible speeds to match your GPU NICs\/DPUs.<\/li>\n<li><strong>Spine-leaf ready<\/strong> \u2014 combine switches into a non-blocking fabric for hundreds to thousands of GPUs.<\/li>\n<li><strong>Validated with RDP compute<\/strong> \u2014 delivered as part of a tested cluster, not a loose box.<\/li>\n<li><strong>On-prem sovereignty<\/strong> \u2014 your fabric, your data path; DPDP-friendly, air-gappable.<\/li>\n<li><strong>Make-in-India OEM<\/strong> \u2014 predictable INR pricing, GST tax invoice (HSN 8517), pan-India onsite support, GeM-procurable.<\/li>\n<\/ul>\n<h3>Where it fits<\/h3>\n<ul>\n<li><strong>GPU cluster fabric (primary):<\/strong> the interconnect between GPU servers that carries distributed-training collectives.<\/li>\n<li><strong>Storage fabric:<\/strong> high-bandwidth path between GPU nodes and the NVMe parallel-FS (GPUDirect).<\/li>\n<li><strong>Scale-out spine:<\/strong> a leaf or spine in a non-blocking topology for large clusters.<\/li>\n<li><strong>HPC interconnect:<\/strong> tightly-coupled MPI and AI collectives.<\/li>\n<\/ul>\n<h3>How it works<\/h3>\n<p>The switch forms a leaf\/spine fabric with RDMA over InfiniBand; GPU NICs\/DPUs connect at 800G \/ 400G. In-network compute (SHARP) offloads reductions, and adaptive routing avoids hotspots \u2014 so all-reduce traffic, which dominates multi-node training, completes fast. RDP sizes the topology (oversubscription, spine count) to your cluster. <em>Honest note: real scaling efficiency depends on topology, collective library and model \u2014 we validate it on your cluster, not just a peak port number.<\/em><\/p>\n<h3>Industry use cases<\/h3>\n<ul>\n<li><strong>AI\/ML &amp; foundation-model teams<\/strong> \u2014 the fabric that lets training scale across nodes.<\/li>\n<li><strong>Government &amp; national labs<\/strong> \u2014 sovereign HPC\/AI interconnect.<\/li>\n<li><strong>Neocloud \/ AI providers<\/strong> \u2014 the fabric layer of a GPU cloud.<\/li>\n<li><strong>BFSI &amp; research<\/strong> \u2014 low-latency clusters for training and HPC.<\/li>\n<li><strong>Telecom<\/strong> \u2014 high-bandwidth AI fabric at core\/edge.<\/li>\n<li><strong>Manufacturing &amp; energy<\/strong> \u2014 HPC + AI cluster interconnect.<\/li>\n<\/ul>\n<h3>Performance \u2014 and how to be sure<\/h3>\n<p>We don&#8217;t publish inflated peak numbers. The honest picture: 102 Tb\/s non-blocking at &lt;600 ns port-to-port is sized to carry cluster collectives without becoming the bottleneck. <strong>Want certainty? Request a free scaling test of your training job across nodes on this fabric before you buy<\/strong>; we&#8217;ll send back real all-reduce bandwidth and scaling efficiency.<\/p>\n<h3>Series &amp; scale path<\/h3>\n<ul>\n<li><strong>DRACO<\/strong> (flagship fabric tier) \u2014 <em>this<\/em>.<\/li>\n<li><strong>Speed ladder:<\/strong> 200G Ethernet \u2192 400G NDR \u2192 800G XDR InfiniBand.<\/li>\n<li><strong>Scale:<\/strong> single switch \u2192 leaf-spine fabric \u2192 full non-blocking fabric for thousands of GPUs (see RDP AI Networking fabric bundles).<\/li>\n<\/ul>\n<h3>On-prem vs cloud \u2014 the TCO case<\/h3>\n<p>An owned fabric is a one-time asset that removes the per-node networking premium of cloud GPU clusters and keeps the data path in-house. RDP pricing is fixed in INR with a GST input-credit-eligible invoice \u2014 ask for a <strong>fabric design and TCO<\/strong> for your cluster.<\/p>\n<h3>Software &amp; integration<\/h3>\n<p>Integrates with the AI stack: the InfiniBand subnet manager, NCCL with SHARP, and standard MPI, with fabric monitoring and telemetry. Validated with RDP GPU servers, rack-scale systems and storage.<\/p>\n<h3>Power, cooling &amp; rack integration<\/h3>\n<p>A 1U switch with redundant PSUs and front-to-back airflow \u2014 specify rack power, cooling and cable plant (DAC\/AOC\/transceivers). <em>(Exact power draw, BTU and cabling confirmed on the build sheet.)<\/em> Full out-of-band management.<\/p>\n<h3>Deployment, warranty &amp; support<\/h3>\n<ul>\n<li><strong>Made to order<\/strong>, configured, cabled and validated in India as part of the cluster; lead time confirmed at quote.<\/li>\n<li><strong>In the box:<\/strong> switch, rails, power cables, and the configured fabric profile; transceivers\/cables scoped at quote.<\/li>\n<li><strong>Onsite warranty + AMC<\/strong> with pan-India coverage and an RMA\/escalation path <em>(exact terms confirmed at quote)<\/em>.<\/li>\n<\/ul>\n<h3>Why RDP<\/h3>\n<p>14 years of Make-in-India infrastructure and <strong>300,000+ devices shipped<\/strong>. Indian OEM, INR pricing, GST tax invoice, pan-India onsite engineers, GeM availability, and DPDP \/ sovereign-AI-ready deployment.<\/p>\n<h3>Buy with confidence<\/h3>\n<p>This is an AI cluster fabric switch, made to order \u2014 <strong>talk to an RDP solution architect<\/strong>, design the topology and oversubscription for your cluster, get a fabric TCO, and <strong>run a scaling test before you commit.<\/strong> Request a quote to begin.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>64\u00d7 800G OSFP (XDR) \u00b7 102 Tb\/s switching \u00b7 InfiniBand XDR (Quantum-X800 class) \u00b7 &lt;600 ns port-to-port \u00b7 1U<\/p>\n","protected":false},"featured_media":1952,"comment_status":"open","ping_status":"closed","template":"","meta":{"_yoast_wpseo_title":"","_yoast_wpseo_metadesc":"","rank_math_title":"DRACO 64-Port 800G InfiniBand Switch \u2014 102 Tb\/s, InfiniBand XDR (Quantum-X800 class) | RDP GPU Mart","rank_math_description":"On-prem AI cluster fabric \u2014 64\u00d7 800G OSFP (XDR), 102 Tb\/s non-blocking, InfiniBand XDR (Quantum-X800 class), <600 ns port-to-port. Wire your GPU cluster at full bandwidth. Make-in-India, GST invoice. Request a quote.","_hermes_jsonld":""},"product_brand":[],"product_cat":[23],"product_tag":[],"class_list":["post-1832","product","type-product","status-publish","has-post-thumbnail","product_cat-ai-networking","pa_form-factor-1u","pa_series-draco","first","instock","taxable","shipping-taxable","product-type-external"],"_links":{"self":[{"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/product\/1832","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/product"}],"about":[{"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/types\/product"}],"replies":[{"embeddable":true,"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/comments?post=1832"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/media\/1952"}],"wp:attachment":[{"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/media?parent=1832"}],"wp:term":[{"taxonomy":"product_brand","embeddable":true,"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/product_brand?post=1832"},{"taxonomy":"product_cat","embeddable":true,"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/product_cat?post=1832"},{"taxonomy":"product_tag","embeddable":true,"href":"https:\/\/rdp.in\/gpu-mart\/wp-json\/wp\/v2\/product_tag?post=1832"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}