Resilience and RO2

100% uptime guarantee that's not smoke and mirrors

Why resilience is important

Connectivity is the life blood of many organisations, and too often resiliency, RO2 and having a second diverse connection is not considered; mostly on the basis of cost and an acceptance of the risks.

However, what would the impact be to your organisation if the primary and only circuit failed?

  • Would this mean sending home 100s of your staff?
  • Would this mean your phone lines stop running?
  • Would this mean losing access to all your cloud business applications?
  • Would this cost you lost sales and customers?
  • Would this leave your brand reputation at risk?
  • How long could you cope without connectivity?

As connectivity is becoming ever more critical to organisations, and as no one can realistically guarantee something to be 100% fault free, resiliency is an absolute priority for most organisations.

Be aware of other carriers promising 100% or 99.99+% uptime on single connections; even if they are leased lines. This is pure marketing smoke and mirrors.

Giganet's 100% SLA for RO2 solutions

At Giganet, we're realistic in that the only way we can offer a 100% SLA is with an RO2 resilient design, and we'll explain what that means here.

What is RO2

RO2

RO2 describes a fully-managed leased line service whose aim is for no single points of failure. This is why we can offer a 100% uptime SLA as we're specifically engineering the design with no-single points of failure, guaranteed.

  • Dual high-availability (HA) Giganet managed Juniper SRX routers operating in Cluster/VRRP HA mode
  • Same fibre-tail provider being used for both primary and secondary routes (as they are in full control and know where their ducts/fibre routes)
  • Different fibre entry points to your premises (where possible)
  • Different fibre routes outside your premises to the exchanges, without 'pinch points' (where possible)
  • Different terminating local exchanges/points of presence (POPs)
  • Different backhaul routes from the local exchanges/POPs to Giganet's core network
  • Different terminating Giganet data centres for the primary and secondary connections
  • Resilient Giganet core network with diverse ring, peering and transit links
  • Full BGP routing to ensure auto failover of customer IP addresses in case of any failure along the path

To guarantee the no-single point of failure, ironically, the same underlying wholesale Ethernet carrier and fibre-tail provider has to be used for both the primary and secondary circuits. This is because they know where their fibre routes, and can ensure no pinch-points. With RO2, you need to work with the same Internet Service Provider (ISP) - in our case this is us - Giganet.

RO2 Updated_Web

Caveats with RO2

Although RO2 solutions can be a silver bullet for many organisations, there are some limitations and caveats worth knowing about.

RO2 is not available across the entire UK, and in particular, remote and rural regions of the UK, or hard to reach locations with single road access/fibre routing to, or where multiple exchanges are not available or the carrier is not located within them, will likely be outside of the scope of where RO2 is available.

Our pre-sales team can find this all out at the time of quotation.

However there are some factors which can only be discovered after a site-survey.

'Pinch Points' may exist with the fibre routing.

Where an RO2 design cannot entirely avoid pinch-points (as defined above), then our 100% SLA reduces to a 99.995% uptime SLA. This is because there are single-points of failure, so we can't guarantee 100% uptime.

Usually pinch-points can be overcome, but this often involves civil engineering and potential for very high install charges.

For example, there may only be one entry/exit point for the fibre circuit into your building (unless you want to spend a few £1,000 on creating a new one). This means that if there were some works at this entry point, then both circuits could be cut. Or there could be 'Pinch points' in the road/pavements outside where for the first few hundred metres, or perhaps more, the fibre is routing down the same fibre core or ducts before splitting off to the different exchanges. If someone is digging in the road, then there's the risk that both circuits are cut.

We divulge any potential caveats after survey

To ensure that you are kept fully informed with any potential caveats, we will advise you if there are any potential 'pinch points' or other possible issues that may jeopardise the 'no single point of failure' target and whether this impacts the SLA.

You may decide to accept this risk, or you may deem RO2 not to be worth it and an alternative solution may be taken. We can advise you of the options.

Alternatives to RO2

If RO2 is out of your budget or not available then there could be alternative options that you may wish to consider. However RO2, where available, is the gold-standard of resilient connectivity and from which we can provide our 100% uptime SLA.

Giganet's Fibre Exchange and Multiple Carrier Options

Giganet has a Fibre Exchange where we interconnect with multiple national fibre and broadband carriers. This enables us to not put all of our eggs in one basket as you would if you went to BT, Virgin or TalkTalk. We can engineer solutions with multiple different carriers for primary, secondary and in some situations tertiary circuit configurations.

1. Diverse carriers + different fibre tail providers - 99.995% uptime SLA

Via the Giganet Fibre Exchange, we can architect a diverse carrier option for the primary and secondary circuits, including a different fibre tail provider. With this option, we can use a carrier that uses one fibre tail provide for the primary circuit (e.g. Openreach tail), and for the secondary circuit we can use a different fibre tail provider (e.g. CityFibre).

This option is the next-best after RO2.

What this means is that we can terminate the primary and secondary circuits onto different carrier's networks, via different physical fibre networks in the local area, via potential diverse backhaul, and also terminate these connections into different Giganet core data centres.

This is almost like RO2, in some cases better (as it's using different fibre infrastructure providers), however unlike RO2, it doesn't come with any guarantee of separacy of fibre routing and each fibre tail provider won't share to the other where they route their fibres.

 

Caveats with diverse carriers

  • 'Pinch points' for the entry/exit into your premise. Both the primary and secondary circuits may enter the building at the same location, however this is usually less likely than if the same fibre tail provider is used.
  • 'Pinch points' between your office and the exchange. The primary and secondary fibre tail provider may be sharing the same ducts for some of its routing.
  • It's unlikely, but it's possible that the same exchange/point of presence is used for both carriers (a fire/flood/major power outage (although backup generators are installed)). Usually we can find this out at the quote stage.
  • Potential for the same backhaul being used for both carriers.
Different Providers_web-06

2. Diverse carriers + same fibre tail providers - 99.99% uptime SLA

What do we mean by "fibre tail provider"? This is the infrastructure provider that owns the physical fibre in the ground, e.g. (BT Plc) Openreach, Virgin Media, Colt, CityFibre, Glide, Gigaclear + more.

How does the "fibre tail provider" differ from the "carrier"? The carrier is the network operator who purchase the fibre tail and connect this to their network before handing it off to Giganet. They are often the carrier located in the (BT Plc) Openreach exchange who have unbundled it, e.g. TalkTalk, Sky, SSE + many more.

Via the Giganet Fibre Exchange, we can architect a diverse carrier option for the primary and secondary circuits. Often, due to which carriers and fibre tail providers that are available at the customer's location, this may involve the same fibre tail carrier (most likely Openreach).

What this means is that we can terminate the primary and secondary circuits onto different carrier's networks within the exchange, and their different hardware and potentially different backhaul and different handoff connection to Giganet, and also terminate these connections into different Giganet core data centres.

 

Caveats with diverse carriers but with the same fibre tail provider

  • 'Pinch points' for the entry/exit into your premise.
  • 'Pinch points' between your office and the exchange.
  • Same exchange is used for both carriers (a fire/flood/major power outage (although backup generators are installed)).
  • Same backhaul is used for both carriers (for instance although we may use TalkTalk Business for the primary circuit, and a Virtual1 for the secondary, they may both be using Virgin Media or SSE for the backhaul).
Same Provider_Web-03

3. Diverse carrier Backup broadband - 99.99% uptime SLA

This is the lowest cost and simplest solution that we recommend for all customers, even where they consider backup not to be important. It's very low cost, and can help provide protection against common issues and maintenance activities. We can use a different broadband carrier to the primary leased line carrier which can help to try and avoid any pinch-points.

 

Caveats with backup broadband

  • 'Pinch points' for the entry/exit into your premise - both fibre and copper could route the same way.
  • 'Pinch points' between your office and the exchange - both fibre and copper could route the same way.
  • Same exchange is used for the primary fibre leased line and the secondary broadband (a fire/flood/major power outage (although backup generators are installed)), however in some cases the broadband can be at a different exchange.
  • Same backhaul is used for both the leased line carrier and the broadband circuit. (This is difficult to ascertain the information for however).
RO2_Backup Broadband-03

4. Backup 4G

This is an option that Giganet is attempting to bring to market in 2019 where we can provide fully managed failover for IP addresses across a primary fibre leased line and a backup 3G/4G modem.

 

Caveats with backup 4G

  • 4G network is not really designed to operate an entire office's data requirements due to the capacity being shared with many 100s or 1000s of mobile users, so speeds and latency may fluctuate. (5G may make this easier in the future).
  • Limited data allowance for 4G, and overage data is charged at £40/GB, so bills could add up, but is this a price worth paying for continued service?
  • Not really possible to failover all applications, in particular realtime apps such as VoIP/ video. The latency and jitter characteristics with 3G/4G could make realtime applications unworkable.
  • Not really possible to provide QoS (quality of service).
  • The 4G mast's backhaul could be routing via the same exchange and or the same backhaul carrier. (There is no real possibility of finding this detail out.)
4G Backup_Web

The competition's answer to resilience

What are others offering with regards to SLAs and resilience?

"Fake 100% uptime SLAs"

A few people come to us each week claiming that other ISPs can offer 100% uptime SLAs, "and why can't we"?

This is pure smoke and mirrors and doesn't accurately reflect the risk and probability of failure.

The 100% SLAs provides you with a false sense of security. If something is quoted as being "100% available", then why would you ever consider a resilient circuit? You wouldn't.

However what happens if the "100% uptime guaranteed" circuit were to fail? You would typically receive a refund of your rental charges against the period of downtime.

Is this a fair representation of the inconvenience caused? Probably not.

Why is 100% uptime unrealistic with a single connection? Engineering works in the road/pavement causing fibre breaks in the ducts, the ISP has maintenance on their network that they do from time-to-time for performance/security, electronic items just break (MTBF figures exist), human error. These are all quite likely things to cause outages. There are then things like fires and floods in the exchanges, which may be unlikely, but these have happened!

Larger ISPs - they limit choice to their own network

Larger ISPs such as the UK's largest two providers typically only leverage their own network for circuits, including backup ones.

This can result in:

  • Reduced choice - as you have to use their network for both primary and secondary circuits, and others may be better value.
  • Increased costs - as they may not have resilient network assets close to you, and therefore have to long-haul them in at additional cost.
  • Smaller fish in their sea - when you need support, how quickly will they answer and help, or will you be passed around call centre workers that don't understand your organisation?