Most offices in Hyderabad that tell us they have internet redundancy have two broadband connections and a man who swaps the cable over. Which is not redundancy. It is a second line and a plan to notice.
The question worth answering before spending anything is a narrow one: when the line goes down, what is supposed to happen, and who has to be awake for it?
The failure you are actually planning for
Line failure is the obvious one, and the rarest. The ones that take an office offline are duller:
- The ISP is up, but routing to one particular service is not — your accounting portal is unreachable while everything else works.
- The link is up and passing traffic at a fraction of its speed.
- The fibre is cut somewhere between the exchange and the building, and the ETA is "today".
- The router reboots itself in the middle of the working day and nobody can say why.
A second connection helps with all of these only if something automatically decides to use it. Without that, the second line is insurance you still have to claim on by hand.
Option one: two broadband lines with automatic failover
Two connections from different providers, terminating on a router or firewall that watches both and moves traffic when one stops answering.
What it gets you. Staff keep browsing, email keeps working, and the switch happens in seconds rather than whenever somebody notices. For most offices this is the honest answer.
Where it falls short. Your public IP changes when the line changes. Anything that depends on a fixed address breaks at exactly the moment you needed it not to: a VPN that remote staff connect to, a camera system published for off-site viewing, a server whitelisted by a bank or a government portal, your own mail server.
Existing sessions drop too. A call in progress on an IP phone ends; a long upload restarts. Failover is not seamless, it is fast.
Who it suits. Offices where the internet is something staff consume rather than something the outside world connects into. Which is most of them.
Option two: BGP, your own addresses, and peering
Here you hold your own IP range and announce it to more than one provider. The address stays yours. If one carrier fails, the same addresses are reachable through the other — no renumber, no reconnect, and sessions that survive the change.
The price of it. This is not a product you buy, it is an arrangement you run. You need your own address space and an AS number, agreements with more than one upstream, routers that speak BGP, and somebody who will notice when an announcement is wrong. Done carelessly, BGP breaks in ways that are much harder to diagnose than a dead cable — and it can break reachability for other people, not only for you.
Who it suits. Organisations the outside world connects into. Anyone hosting a service, running their own mail, publishing cameras or portals, or carrying enough traffic that the route itself matters.
Peering — exchanging traffic directly rather than paying a transit provider to carry it — sits alongside this. It is about where your traffic goes and what it costs, not only about staying up. We cover the mechanics on our Peering and Internet Exchange page, including whether you need your own AS number.
What actually matters
Four things decide this, and price is not among them.
Does anything connect in to you? If the answer is no, dual broadband with failover is almost certainly right, and BGP is an expensive way to solve a problem you do not have.
What breaks when the IP changes? Write the list down. VPNs, remote camera access, whitelisted addresses at banks and portals, DNS records, licence servers. If that list is long, failover alone will not save the day it is needed.
Who is on duty? BGP that nobody monitors is worse than no BGP. A misannouncement is silent from inside the building: your staff see nothing wrong while the rest of the internet cannot reach you.
Are the two lines genuinely independent? Two providers reselling the same backhaul, or entering the building through the same duct, fail together. One JCB takes both. Ask each provider which physical path they use — and get it in writing before you sign.
The middle ground most offices end up at
A single provider, one good router or firewall with a second line for failover, and the few services that genuinely need a stable address moved somewhere that already has one. Hosted mail instead of a mail server in the rack. A cloud VMS, or a VPN that dials out rather than one the world dials into.
Call it a recognition rather than a compromise: that the stable address was the only reason BGP was on the table.
Where we come in
We design and run both, and we will tell you which one your office needs rather than selling the larger of the two. In practice that means:
- Looking at what connects inward before recommending anything.
- Dual-WAN failover on routers and firewalls configured properly, with the failover actually tested rather than assumed.
- Peering assessment, BGP configuration, IX connectivity and multi-ISP routing where the case is there, with monitoring and ongoing support — because an announcement nobody watches is the part that bites.
- Network AMC so somebody is responsible when it does fail, which is the part most redundancy plans leave out.
We have been doing networking and security work in Hyderabad since 2008. Call us and describe what connects into your office, and we will tell you which of these two you need.