Free planning tool

    BIG-IQ to Avi Migration Estimator

    For estates managed centrally through F5 BIG-IQ. How much of the estate you can safely retire instead of migrate, how many parallel streams your change control will actually allow, what it costs to rebuild the management plane on Avi - and how long the whole programme takes.

    Rationalisation modelled from your traffic statistics
    Multi-stream wave plan with a phased business case
    Management plane and decommission costed separately

    We don't ask for and won't accept load balancer configuration through this site. Everything below is calculated in your browser from the numbers you enter.

    The cheapest virtual service is the one you never build

    BIG-IQ gives you something a device-by-device estate never has: evidence. Traffic statistics turn every retention argument from an opinion into a measurement, and every virtual server you retire is one you never build, test, cut over or operate again.

    ~35%

    Typically retired, not migrated

    With twelve months of statistics, a third of a large estate turns out to be idle, orphaned or serving something already switched off.

    0.5 vs 1.2

    Days per virtual server

    Central configuration and templating roughly halve per-virtual-server effort against a device-by-device audit.

    1 stream

    What a central change board buys you

    Parallel streams need parallel change authority. Without it, headcount changes nothing.

    Build your BIG-IQ estate estimate

    The two answers that move the number most are how much traffic history you hold and how change is approved. Everything else is pre-filled with the assumptions we use on a large UK enterprise estate.

    Step 1 of 6 - Your BIG-IQ estate

    What do you actually use BIG-IQ for?

    Central configuration and statistics are the two that change the estimate most.

    Decommission is sequenced per group.

    This is the single biggest lever on the estimate. Statistics are how you prove a virtual server carries no traffic and can be retired instead of migrated.

    Which modules are licensed and in use?

    Migrating a BIG-IQ-managed estate to Avi

    A large centrally managed estate - around 1,800 virtual servers across 420 applications - runs to roughly two years, and that is a programme you fund in phases rather than one you approve in a single business case. Above about 78 weeks we recommend two funded phases with a defined boundary; above 130 weeks, three. Each phase gets its own review point, and the later phases benefit from what the earlier ones learned.

    BIG-IQ makes the migration materially easier, not harder. Central configuration turns discovery into an export instead of a device-by-device audit. Templated configuration means whole groups of virtual servers are instantiated from a handful of patterns rather than built one at a time. And statistics let you prove which virtual servers carry no traffic at all. On the estates we model, that is the difference between roughly 1.2 days of effort per virtual server and about half a day.

    What replaces BIG-IQ itself

    The Avi Controller. Central management, analytics, RBAC and configuration all live in the controller cluster rather than a separate management product. The work is not installing a replacement - it is rebuilding the specific capabilities you actually rely on. The estimator costs those individually, because most estates use a much shorter list than the BIG-IQ feature matrix suggests, and paying to rebuild things nobody uses is a common way to lose a quarter.

    Change control is the real constraint

    Parallel migration streams need parallel change authority. If every business unit approves its own changes, several streams can run at once and the programme compresses dramatically. If one central board approves everything, the programme is capped at a single stream no matter how many engineers you fund - and on a large estate that is the difference between two years and four.

    Typical estate benchmarks

    Estate profileStreamsTypical elapsed time
    800 virtual servers, 200 apps, 12 months of statistics2 streams60–70 weeks
    1,800 virtual servers, 420 apps, partial statistics3 streams100–108 weeks, run as two phases
    1,800 virtual servers, 420 apps, no statistics3 streams115+ weeks, retention drops to ~86%
    1,800 virtual servers, 420 apps, one central change board1 stream220+ weeks - the change board is the programme

    Managing BIG-IP device by device rather than through BIG-IQ? Use the F5 BIG-IP to Avi migration estimator instead.

    BIG-IQ to Avi: common questions

    What replaces BIG-IQ when we move to Avi?

    The Avi Controller does. Central management, analytics, RBAC and configuration are all built into the controller cluster rather than sitting in a separate management product. The work is not installing a replacement - it is rebuilding the specific BIG-IQ capabilities you actually rely on, which is usually a much shorter list than the feature matrix suggests.

    Does BIG-IQ make the migration easier or harder?

    Easier, materially. Central configuration means discovery is an export rather than a device-by-device audit, templated configuration means large groups of virtual servers can be instantiated from a handful of patterns, and statistics let you prove which virtual servers carry no traffic. On the estates we model, that is the difference between about 1.2 days of effort per virtual server and about half a day.

    How much of a BIG-IQ-managed estate can we safely retire?

    More than most teams expect. With twelve months of statistics, a quarter to a third of virtual servers typically turn out to be idle, orphaned or serving decommissioned applications. Every one you retire is a virtual service you never build, never test, never cut over and never operate again.

    We have no statistics. What should we do first?

    Turn statistics collection on and wait a quarter before the build starts. On a large estate that fortnight of setup is typically worth several months of programme time, because it converts every retention argument from an opinion into a measurement.

    Can we run several business units in parallel?

    Only if each has its own change authority. Parallel streams need parallel approvals, and a single central change board caps the whole programme at one stream regardless of how many people you staff. That constraint, not headcount, is usually what makes a large estate take two years instead of one.

    How do BIG-IQ tenants and roles map to Avi?

    Avi tenants map cleanly to business-unit boundaries, with roles and RBAC defined per tenant on the controller. The design work is agreeing which boundaries are real isolation requirements and which are administrative convenience carried over from BIG-IQ.

    What happens to our VE licences?

    They unwind as devices are retired, which is why the estimator sequences decommission per device group rather than leaving it all to the end. Getting licence release into the wave plan is often where the business case is actually realised.

    Should a two-year programme be split into phases?

    Yes. Above roughly 78 weeks we recommend running the work as two funded phases with a defined boundary, and above 130 weeks, three. It keeps the business case reviewable, and it lets the second phase benefit from what the first one learned.

    Turn the estimate into a funded plan

    Send us a BIG-IQ inventory export and we'll tell you what can be retired, what converts, and where the phase boundaries should fall. No charge, no obligation.