Weighted, load-based routing
Traffic goes to the lowest-workload node, not round-robin. Imbalances correct automatically, with no client-side changes.
Load Balancing Built for Object Storage
A storage cluster is only as fast as your ability to use all of it. HyperBalance spreads S3 traffic evenly across your nodes so you get the throughput you paid for — with no hot spots, no idle capacity, and no client-side changes as the cluster grows.
Load Balancing Without a Load Balancer Tier
HyperBalance Fusion is a software-defined load balancer built into HyperStore and enabled with a license. No separate tier, no external proxies, no third-party global load balancing.
It routes through DNS by default: weighted, load-aware decisions send each client to the optimal node while the load balancer stays off the data path, so throughput is never capped by a proxy tier. When DNS alone can’t keep up — AI training bursts, backup windows, clients that pin a single connection — the relay feature engages in real time, redistributing only the affected sessions.
Off-path efficiency, with a safety net for the workloads that need one.

Traffic goes to the lowest-workload node, not round-robin. Imbalances correct automatically, with no client-side changes.
HTTP-level checks reflect live HyperStore service status per node, a readiness distinction generic TCP checks can’t make.
Active-active across datacenters, traffic kept local to the originating site, health-aware failover. No external GSLB license.
Certificates generated and terminated centrally — nothing to rotate on individual nodes.
Node additions, removals, and failures republish automatically.
Per-tenant VLAN tagging, with internal node addresses kept private.

An Independent Tier. Full Control of the Traffic Layer.
When load balancing needs to be configured, tuned, and operated independently of storage, or policy requires that separation, HyperBalance gives you a dedicated tier with the full range of Layer 4 and Layer 7 routing control.
It runs as software on industry-standard servers, sitting inline on the data path and separate from your HyperStore nodes. It’s also the answer for the largest deployments: where Fusion is validated to 80 nodes, HyperBalance is built beyond that.
Routing flexibility across both layers, with the configurability enterprise environments expect.
Inline balancing, GSLB, or both — matched to your topology.
Deployed as an HA pair for uninterrupted service.
No proprietary platform required; two servers for an HA pair.
Full segmentation for multi-tenant and service provider environments.
A dedicated tier keeps load balancing fully independent of storage — and scales past Fusion’s 80-node ceiling.
| HyperBalance Fusion | Hyperbalance | |
| Deployment | Software feature of HyperStore, runs on every node | Software on dedicated industry-standard servers |
| Dedicated servers required | None | Yes — two for an HA pair |
| Load balancer type | Storage-native, DNS-based with selective relay | Enterprise load balancer, Layer 4 + Layer 7 |
| Data path | Off the data path by default | Inline — on the data path |
| Performance | Throughput scales with the cluster | Depends on server sizing |
| Health checks | S3/HTTP protocol-level, integrated with HyperStore | Available, not integrated with HyperStore |
| Load-based routing | Weighted and load-aware, integrated | Weight-based available, not integrated |
| Multi-DC support | Native GSLB, active-active | Inline balancing, GSLB, or both |
| High availability | DNS-based, active-active or failover | Active-passive HA pair |
| VLAN/multi-tenancy | Yes | Yes |
| Validated scale | Up to 80 nodes | Largest clusters |
| Licensing | Licensed per HyperStore node | Separate product |
Choose Fusion when you want an integrated solution, don’t need the separation of an independent tier, and want throughput that scales with the cluster.
Choose HyperBalance when you need the full enterprise feature set, dedicated separation from storage, clusters beyond 80 nodes, or complex multi-vendor infrastructure.