We're about to colo in a remote facility. The services provided are time-critical and very high traffic, but do not require a large IP allocation. I have an AS from ARIN, but have nowhere near the IP utilization requirements to justify a full /24 delegation, much less a direct /22 ARIN assignment. The core question here is how can we implement redundancy and load-sharing if we cannot justify a /24 delegation from the ISP? Even is we could justify the delegation, I've heard that most ISP's will not allow you to advertise a different provider's IP delegation. We're only talking about 35 - 50 IPs at the site.
If that is true how do most organizations handle service-provider redundancy and load-sharing with small IP spaces? We will be pushing a very high volume of time critical traffic and want service-provider redundancy. So what does that leave me with? I must assume other people have these type of requirements. We have a very high-bandwidth application but as a startup I cannot afford 6500 series gear. My best ROI comes from 3750s right now, but they do not support assigning a NAT for each providers IP space to each virtual IP. Is the only option to upgrade to 4500 or 6500's in order to NAT at very high traffic volumes (few hundred Mbs)? Is there a solution I'm not aware of?