Data Mobility Switch - Features and Benefits

The Data Mobility Switch (DMS) is purpose-built to accelerate all IP traffic traversing Big Traffic WANs.  The DMS reduces multi-gigabit traffic by as much as 5x or more at traffic speeds of up to 10Gbps. The DMS is ideal for accelerating critical workflows such as replication, long distance ‘live’ migrations, massive file transfers, and Big Data.

Tailored for Data Centers

A core component of the DMS is its massive scale, fully non-blocking switch fabric. The switch fabric allows the DMS to be deployed inline or out-of-path in data centers without the DMS becoming a bottleneck, even in environments where routing asymmetries may occur.

The DMS can accelerate up to 10Gbps of traffic even during sustained data bursts of well over 10Gbps. Traffic that matches configured policies will be processed for acceleration, while the rest is bridged through the DMS in less than 2 micro-seconds. In contrast, traditional WAN optimization products (given their off-the-shelf server-based design) cannot be deployed inline in data centers, and require expensive load balancing equipment to redirect traffic to them.

The switch fabric also enables the DMS to support fail-to-wire functions in both inline and out-of-path deployments. This is a major advantage over traditional WAN optimization solutions that offer no native fail-to-wire capabilities for data center deployments.

True Transparency

Makers of WAN optimization products have traditionally defined network transparency as the ability to maintain IP addresses and TCP ports across the WAN for traffic being accelerated. Although this makes sense for branch WAN scenarios, inter-data center WAN optimization solutions must also incorporate the following features to be considered “transparent”:

• Low port-to-port latencies: Port-to-port latencies for inter-data center WAN optimization solutions must be in the low 10s of microseconds. This is important because applications traversing inter-data center WANs tend to be extremely latency sensitive and will deteriorate in performance if the end-to-end transaction latency goes up even slightly.
• End-to-end packet boundary preservation: Inter-data center WAN optimization solutions must guarantee that packet sizes and counts are preserved end-to-end across the WAN. This is important because some applications traversing inter-data center WANs break when packets are transparently re-segmented or re-sized.

Not only can the DMS maintain IP addresses and TCP ports across the WAN for all inter-data center traffic, DMS port-to-port latencies average 50 micro-seconds and the system guarantees end-to-end packet boundary preservation. The DMS is the only transparent WAN optimization solution for data centers.

Scalable, Sustained Data Reduction

The DMS is the only WAN optimization solution capable of deduplicating WAN traffic at multi-Gigabit rates, reducing critical application traffic by 5x or more. Whether ingress speeds are 1Gbps or 10Gbps, the DMS delivers equally impressive reduction ratios. Traditional WAN optimization solutions sacrifice reduction efficacy to keep up with increasing ingress rates. In fact these solutions may perform no deduplication whenever ingress rates exceed a few hundreds of Mbps.

Fat, Aggressive Per-Flow Acceleration

The DMS supports 1Gbps+ per-flow speeds for throughput-intensive, bursty traffic. Applications can achieve these speeds over any WAN at any distance, making flow throughput independent of WAN latency or loss characteristics.

The DMS effectively shields applications from the impact of network congestion and contention by managing WAN resource allocation and by using a unique congestion-control algorithm that has been designed for highly aggressive flows. Be it 10 flows on a 10Gbps link or 100,000 flows on 1Gbps link, the DMS ensures optimal WAN utilization at all times.