BYOL reprices OCI database services by roughly three quarters using licenses you already own, one Enterprise Edition processor covering two OCPUs, on conditions the audit machine can check: support active, options matched, one deployment per license. License included is the flexible answer, elasticity, bundled options, clean exits, support independence, and this session prices both honestly, workload by workload. Module 2's shelfware ledger becomes a cloud currency inventory, the harvest sources get ranked, the third party interaction gets mapped, and the worked example lands a $376K a year saving on eight licenses that were gathering dust.
The presenter in this session is an AI generated avatar. The curriculum and guidance are real, produced by Redress Compliance analysts from our consulting engagements and market network.
A taught session with three knowledge checks: eight shelf licenses converted into a 16 OCPU workload's 75 percent rate cut, the third party supported licenses discovered BYOL ineligible and the portfolio interaction mapped, and the undecommissioned fallback server diagnosed as the compliance gap it is. It closes with one workload priced both ways: license included at $620K against BYOL at $244K all in, and the provisos that keep the saving from becoming a finding.
The full narration of this session, section by section, for reading and reference.
Welcome back, session twenty seven of forty. Last session you learned the OCI commercial machine: the credit pool, the commit, the expiry clause, the forecast. Today, the seam where that machine meets everything from the first five modules: your perpetual licenses, and what they're worth in the cloud. The construct is called BYOL, bring your own license, and here's the headline before any detail: applied correctly, it reprices OCI database services by roughly three quarters. Not a discount you negotiate; a rate you qualify for, using licenses that may currently be sitting on your shelf generating nothing but a support bill. If that sounds like module two's shelfware ledger suddenly acquiring a purpose, it is exactly that, and the estates that ran the baseline audit from session sixteen and the entitlement library from session twenty five arrive at this session holding a currency they didn't know they had. But BYOL is a licensing program, which means it has conditions, and the conditions are pure course material: support must stay active, options must match, and one license covers one deployment, with the audit machine from module five perfectly capable of checking all three. Today: what BYOL is, the conversion arithmetic, what license included buys for its higher rate, how to choose per workload, where the licenses come from, and the rules that bite. Let's price the seam.
Five takeaways. One, you'll define BYOL precisely: what the program grants, what it requires, and what it actually costs once the support stream is counted, because the honest total is the only number worth comparing. Two, you'll run the conversion: processor licenses and named user licenses translated into OCI service capacity at the published ratios, the arithmetic that turns an entitlement library into a cloud budget. Three, you'll price license included fairly: the higher hourly rate buys real things, elasticity, bundled options, clean exits, support independence, and for some workloads that bundle is the bargain, not the extravagance. Four, you'll harvest the estate: module two's surpluses, decommission gains, consolidation dividends, all convertible into avoided cloud spend, which is the most satisfying arbitrage this course contains. And five, you'll respect the rules: active support, no double duty, matching options, clean allocation records, each one checkable in an audit, each one trivially easy to follow with session twenty five's change gates and trivially easy to violate without them. The running theme, one more time: the licenses you own are assets with conditions, and knowing both halves is the job. The stakes, next.
Four numbers. Seventy five percent, roughly: the hourly rate reduction of BYOL against license included on comparable OCI database shapes. It is the biggest single lever in OCI economics, bigger than any commit discount session twenty six discussed, and it's available only to estates that can prove their entitlements, which is why the records keep mattering. Two: OCPUs of Enterprise Edition database service covered by one processor license under the BYOL policy. That ratio anchors every calculation today, and its published cousins cover Standard Edition and the named user metrics. One: deployment per license. The license doing BYOL duty in the cloud cannot simultaneously cover the on premises server it came from, outside genuine transition windows, and the fallback server nobody decommissioned is this module's signature audit finding. And twenty two percent: the support stream that must stay active on every license brought. Module four's annuity turns out to be the BYOL entry ticket, which rewires some of module four's own termination arithmetic, and we'll meet that interaction head on in the knowledge checks. One session, one decision, priced from both sides: bring the license, or rent it by the hour. Both answers are right, for different workloads, and the session's job is teaching you which is which. The construct, first.
What BYOL actually is: a pricing program, not a license transformation. Five facts. The grant: existing perpetual licenses with active support may be applied to eligible OCI services, database services above all, at the BYOL rate instead of the license included rate. Your license stays exactly what it was; the cloud service just charges you less because you brought it. What you keep: the perpetual license itself, its support stream, and, importantly, the right to move it back on premises later. BYOL is reversible in a way subscriptions never are, which makes it a low regret first cloud move for cautious estates. What you must hold, and this is the conditions list the whole session enforces: active support on every license brought, the matching edition, and matching options for every feature the cloud service actually uses. The entitlement rules you learned in modules one and five travel to the cloud intact. Where it applies: OCI database services including Autonomous, the Exadata cloud services, and some middleware, with each service's terms naming eligible licenses and conversion ratios. And why Oracle offers it, worth understanding because aligned incentives are rare: BYOL keeps the support annuity alive and makes OCI dramatically cheaper than rivals for existing Oracle customers. Your interest and theirs point the same direction, and the pricing shows it. Read BYOL as module two taught you to read everything: not a discount, a repricing of assets you already own, with conditions. The arithmetic, next.
The conversion table, the five rows that decide the money. Row one, the anchor: one Enterprise Edition processor license covers two OCPUs of Enterprise Edition database service. Every BYOL business case starts from that ratio, so hold it. Eight licenses, sixteen OCPUs. Twenty licenses, forty. Row two, Standard Edition Two: one processor license covers up to four OCPUs of the Standard Edition service, within SE's usual limits. Cheaper licenses stretching further, consistent with everything session three taught about the editions. Row three, Named User Plus estates: the same services are reachable, per the published NUP conversions, and the minimum user counts from session two still apply. NUP arithmetic follows you to the cloud; it does not retire. Row four, options: Autonomous and the feature rich services map to options entitlements. If the service uses Partitioning, you bring Partitioning. If it uses advanced security features, you bring that option. The matching principle from session twenty three, now applied at deployment time instead of audit time, which is infinitely cheaper. And row five, the disqualifier: licenses without active support convert to nothing. Lapsed support means BYOL ineligibility, full stop, and we'll see what that does to third party support decisions shortly. One instruction with this table: run the conversions against your real entitlement library from session twenty five, not against memory. The ratios are published; the entitlements are yours to prove. And now, the fun version of the arithmetic. Knowledge check one.
Knowledge check one. Module two's baseline audit found eight Enterprise Edition processor licenses sitting on the shelf, support still being paid, covering nothing. A new sixteen OCPU OCI database workload is planned. What are those shelf licenses worth? A, nothing, shelfware is sunk cost. B, the full BYOL saving: eight processor licenses cover sixteen OCPUs, converting the workload from license included to BYOL rates, roughly seventy five percent off the database service cost. C, whatever Oracle will refund for them. Or D, they only have value if the workload stays on premises. Pause here. Run the conversion: eight licenses, times two.
The answer is B, and the arithmetic sings. Eight Enterprise Edition processor licenses, at two OCPUs each, cover exactly the sixteen OCPUs the workload needs. Applied as BYOL, they convert the service from license included rates to BYOL rates, roughly three quarters off the hourly cost, every hour, for as long as the workload runs. On a production database this size, that's six figures a year of avoided credit burn, purchased with licenses that were, per module two's ledger, generating nothing but their twenty two percent support bill. Now notice the compounding, because it's the best part: the support stream that made this shelfware painful is precisely what kept it BYOL eligible. Which changes the module four analysis for these licenses, sometimes decisively: a surplus license with a cloud destination is no longer shelfware, it's prepaid cloud licensing, and its support is no longer waste, it's the BYOL entry fee, now generating a return of several times its cost. This is why the baseline audit and the entitlement library keep paying dividends: you cannot harvest what you haven't counted. The wrong answers, briefly. C misreads the model: perpetual licenses aren't refundable, and their value is realized through use, exactly as here. D was true until OCI existed; the entire point of the program is that it isn't true anymore. A is module two's original sin, sunk cost reasoning walking past real money. The honest caveats go in the plan, not against it: support stays active, no double covering the old server, options checked. All three are coming. But the value is real, large, and sitting on the shelf. What the other rate buys, next.
License included, argued properly, because it's not the expensive option, it's the flexible option, and for some workloads flexibility is the bargain. What the higher rate buys, five things. No entitlements needed: the license lives inside the hourly rate. No conversion math, no support prerequisite, no allocation records, no entitlement audit trail. For an estate with no spare licenses, this is simply the price of the service, and it's an honest one. Perfect elasticity: scale from two OCPUs to twenty for month end and back again, and the licensing scales with the meter. BYOL capacity is bounded by the licenses you brought; license included capacity is bounded by your budget. For genuinely elastic workloads that difference is decisive. Options included: many license included services bundle options and management packs that would each need separate entitlements under BYOL. A workload using three options might find the bundle cheaper than the entitlement shopping list. Price the true totals, not the headline rates. A clean exit: stop the service, stop the cost. No stranded licenses, no support obligation continuing, no redeployment question. Pure subscription economics, and for experiments and short lived systems, exactly right. And support independence, the quiet one: license included workloads need no on premises support stream at all, which matters enormously for estates running module four's reduction program, because it breaks the chain between cloud plans and the annuity. Which sets up the collision. Knowledge check two.
Knowledge check two. The estate moved its legacy databases to third party support in module four, saving half the fee. Those old Enterprise Edition licenses would cover a new OCI workload nicely. Can they BYOL? A, yes, a perpetual license is a perpetual license. B, no: BYOL requires active Oracle support on the licenses brought, and third party support does not qualify, so this workload prices license included, or other licenses are found. C, yes, if the third party provider approves it. Or D, no, and the third party move was therefore a mistake. Pause here. What did session nineteen say Oracle support uniquely provides?
The answer is B, and the terms are explicit: BYOL licenses must carry active Oracle support, and third party support, whatever its considerable virtues, is not Oracle support. The licenses remain perfectly valid perpetual licenses, exactly as session nineteen taught, but they're ineligible for the BYOL rate. So the workload prices license included, or the estate brings different, supported licenses. That's the answer; the interesting part is what it teaches about portfolio decisions. Session nineteen's fit profile decided support strategy per workload. This session adds a column: cloud destination. An estate planning OCI migrations should check BYOL eligibility before routing license pools to third party support, because the moves interact. The termination that saves fifty percent on a frozen estate also strips its licenses of cloud currency. Sometimes the trade is still right: a genuinely frozen system with no cloud future loses nothing by losing eligibility it would never use. Sometimes it's wrong: a license pool with a migration eighteen months out may earn its Oracle support keep several times over through BYOL savings. That's arithmetic, run per workload, and you have all the tools. The wrong answers: A ignores that the rate is conditional even though the license is perpetual, the distinction this whole session lives on. C misunderstands authorship; Oracle writes OCI pricing terms, and the provider has no vote in them. And D overcorrects: the third party decision was made on that estate's fit profile, presumably correctly, and discovering a boundary is not discovering a mistake. Portfolio decisions interact; the estates that map interactions before moving win both moves. Choosing per workload, next.
Choosing between BYOL and license included, workload by workload, five factors. Factor one, do the licenses exist? Surpluses, harvestable shelfware, licenses freed by decommissions. If the answer is no spare licenses and no appetite to buy, license included, decision done, no shame in it. Factor two, how long will it run? Steady production for years rewards BYOL's lower rate handsomely; the saving compounds every month. A three month burst barely repays the entitlement paperwork, let alone a support commitment. Duration is destiny in this arithmetic. Factor three, how elastic is it? Fixed, predictable capacity suits BYOL's license bounded model perfectly. A workload that swings from four OCPUs to forty on seasonal peaks suits the rate that licenses itself as it scales. Trying to BYOL an elastic workload means owning licenses for the peak, which is just cloud shaped shelfware. Factor four, what's the support strategy? BYOL chains its licenses to active Oracle support, indefinitely. If module four's program wants that stream reduced or gone, license included buys the freedom, and the freedom has a price you can now calculate. And factor five, what do the options cost? Feature hungry workloads may need three or four options entitlements under BYOL that license included simply bundles. Compare true totals, always. Where do real estates land? Mixed, nearly always: BYOL on the steady core where licenses exist, license included on the elastic edge. The same portfolio shape as every honest answer this course has given. Where the licenses come from, next.
Where BYOL licenses actually come from, five sources, in cost order. Source one, the shelfware ledger: module two's surpluses, already owned, already supported, already painful. The first place to look and the cheapest currency in the estate, surfaced automatically by session twenty five's position, entitlements minus deployments. This is the harvest the whole course has been quietly stocking. Source two, decommissioned systems: every on premises retirement frees licenses, and, elegantly, the migration itself is a source. The database moving to OCI releases the very licenses it was consuming on premises, which can then cover its cloud incarnation as BYOL. The licenses follow the workload, if the cutover is clean, and we'll test exactly that cleanliness in the final check. Source three, consolidation gains: session four's containment and right sizing work frees processor licenses from over provisioned clusters, each one worth two OCPUs of cloud capacity. Infrastructure discipline turning directly into cloud budget. Source four, the restructure pile: session eighteen's set restructures sometimes leave right sized surpluses behind. Check them against the cloud roadmap before any termination notice goes out, per the interaction lesson from knowledge check two. And source five, purchase, deliberately last: sometimes buying perpetual licenses to BYOL genuinely beats license included on multi year total cost, support included. It happens, especially on big steady workloads. Run it honestly rather than assuming either direction. The rules that police all of this, next.
The rules that bite, because BYOL's conditions are all checkable, and module five taught you exactly who checks. Rule one, support stays active: lapse the support and BYOL eligibility lapses with it, converting the workload to license included rates, in the ugliest tellings backdated. The support renewal for BYOL licenses is not an optional line in module four's program; diary it in the session twenty five calendar with everything else. Rule two, no double duty: a license covering an OCI service cannot simultaneously cover its old on premises home. Transition allowances cover genuine migration windows; a permanent warm fallback is not a transition, it's a finding. We test this in sixty seconds. Rule three, options must match: the cloud service's features in use map to options you own, the session twenty three analysis applied at deployment time, when fixing a mismatch costs a configuration change instead of a settlement. Rule four, the record keeping: which licenses are allocated to which OCI service, recorded in the entitlement library at deployment time, the cloud column of session twenty five's position. Ten minutes per deployment, and the difference between a boring audit and an expensive one. And rule five, the audit reality: BYOL claims are verifiable against entitlements, GLAS knows the ratios as well as you do now, and clean allocation records make the whole conversation a formality. None of these rules is hard to follow. All of them are easy to forget. The change gates catch every one, if OCI deployments route through them. The classic failure, tested. Knowledge check three.
Knowledge check three. A migration moves a production database to OCI under BYOL, using the licenses from its on premises server, exactly as the harvest slide suggested. Six months later, the old server is still running the same database, kept as an unofficial fallback that nobody quite got around to decommissioning. What is the position? A, fine, the licenses cover both, they always did. B, a compliance gap: the licenses cover one deployment, the transition allowance has passed, and the fallback is unlicensed, so decommission it or license it, now, deliberately. C, fine as long as the fallback is rarely used. Or D, a problem only if Oracle finds out. Pause here. How many deployments, how many licenses, and what did the migration plan promise?
The answer is B, and you get there the way this course always gets there: by counting. One set of licenses. Two running deployments: the OCI service consuming the licenses as BYOL, and the old server still running the same database as a comfort blanket. The transition allowance that legitimately covered the migration window closed months ago. So the on premises copy is now an unlicensed Enterprise Edition deployment, which is session twenty three's most straightforward finding family, discoverable by any collection script, priced at list plus back support in any audit report. The fix is a decision made now, on your terms, per session twenty five's self assessment logic. Either decommission the fallback properly, uninstall, document the retirement, file the evidence in the layer where session twenty three's rebuttals live. Or license it deliberately, because the business genuinely wants a standby, in which case session twenty three's DR postures tell you exactly what that costs. What the position cannot remain is drift, because drift compounds: today it's one comfort blanket server; three years from now it's a finding with back support attached and a story nobody remembers. The wrong answers: C confuses frequency with licensing; installed and running is the test, settled in module one, and a rarely used deployment is a deployment. A ignores the allocation: the licenses were counted and recorded against the OCI service; entitlement math doesn't care about intentions. And D, once more with feeling: the position is the position whether anyone is looking, the audit cycle reaches every estate, and treating compliance as a detection game is how you end up funding other people's settlements. The prevention is one line in the migration checklist: cutover includes decommission. Put it in the change gate and this check never happens to you. The full pricing, next.
One workload, priced both ways, the session in six rows. The workload: a sixteen OCPU Enterprise Edition database, steady production, running for years. License included: roughly six hundred twenty thousand dollars a year of credit burn at the license included rate. No entitlements needed, options bundled, clean exit anytime. An honest price for total flexibility. BYOL: the eight shelf licenses from knowledge check one, converted at two OCPUs each, cover the workload exactly. Credit burn drops to roughly one hundred sixty thousand at the BYOL rate. Add the support obligation, about eighty four thousand a year on those licenses, which, note, was already being paid while they sat on the shelf doing nothing. Confirm the Partitioning entitlements against the library for the features in use; this estate holds them. True annual cost: roughly two hundred forty four thousand, against six hundred twenty. BYOL wins by about three hundred seventy six thousand dollars a year, on this workload, for this estate. And the provisos, because they are the difference between the saving and a finding: support renewals diarized, the allocation recorded in the entitlement library, and the old server decommissioned with evidence filed, per knowledge check three. Now flip the conditions and watch the answer flip: no spare licenses, and license included's six twenty is simply the price. An elastic workload, and owning peak capacity licenses turns BYOL into cloud shaped shelfware. A support exit strategy on those licenses, and the eighty four thousand becomes a chain. Price both. Always. Recap, next.
Session twenty seven in three sentences. One, BYOL reprices OCI services by roughly three quarters using licenses you already own, at published conversion ratios, one Enterprise Edition processor to two OCPUs, on three conditions that never sleep: support active, options matched, one deployment per license. Two, license included is the flexible answer, not the wasteful one: elasticity, bundled options, clean exits, and support independence, winning for bursts, wild scaling, empty shelves, and estates whose module four program wants the annuity gone. Three, the estate answer is mixed and mapped: harvested licenses on the steady core, license included on the elastic edge, allocations recorded in the entitlement library, and the interactions with the support strategy priced before either program moves, because the third party decision and the BYOL decision share the same licenses. Next session completes the OCI commercial toolkit: Support Rewards and OCI negotiation. The twenty five cent per dollar arithmetic that turns cloud consumption into on premises support relief, how it stacks with everything module four taught, negotiating the credit commitments themselves, and the OCI concessions that shrink bills on both sides of the estate at once. The session where the cloud starts paying for the datacenter. Homework first.
Homework, about an hour, and it builds your cloud currency inventory. One, harvest on paper: from session twenty five's position, every surplus license listed with its BYOL conversion value, processors times two OCPUs, at the BYOL rates for the services you'd actually use. That one page is the cloud currency inventory, and its total is usually a pleasant surprise. Two, price the seam twice: the workload you put on the calculator last week gets both columns completed properly, license included versus BYOL, support included, options included, per today's worked example. Whichever way it lands, you now own the method. Three, check the interactions: any license pool currently headed for support termination or third party support gets one question asked first, does the cloud roadmap want these for BYOL? Eighteen months of foresight here is worth real money in both directions. Four, write the cutover line: one sentence added to the migration checklist, cutover includes decommission, evidence filed. Session twenty five's change gate enforces it from there, and knowledge check three never happens in your estate. And five, map the allocations: if BYOL runs today, record which licenses cover which services in the entitlement library, the cloud column of the position. Ten minutes now, or an audit conversation later; the exchange rate on that trade never improves. That's the hour. See you in session twenty eight.
Five reads, all free on redress compliance dot com. First, Oracle BYOL, a comprehensive guide: the program, the ratios, the eligible services, and the rules, at reference depth, today's session as a desk manual. Second, BYOL versus license included on OCI and Exadata Cloud at Customer, a cost comparison guide: today's worked example generalized across service types and shapes. Third, Oracle Database licensing on AWS with BYOL: the same seam running through other people's clouds, which is exactly where session twenty nine goes. Fourth, the Oracle cloud licensing policy: the policy text behind the conversion ratios, worth reading in the original the same way session seventeen read the support policies. And fifth, optimizing your Oracle license footprint before renewal: the harvest worked as part of the wider estate program, shelfware to cloud currency included. That's session twenty seven. The licenses met the meter, the shelf turned out to be a wallet, and the rules came along exactly as they always do: support active, options matched, one deployment each, records kept. Next session, the arithmetic gets friendlier still: Support Rewards, where OCI consumption starts paying down the on premises support bill, and the negotiation that makes the most of it. See you there.