iblync · Evaluation guide
How to evaluate a data replication platform: a scorecard for Saudi buyers
A weighted scorecard, a test plan and the evidence to request from every vendor, so your choice of replication platform stands up in procurement review.
Evaluation guide Published 7 min read
- Built and supported by Interkey An Interkey product, delivered and supported by Interkey
- One platform for migration and continuous sync The move and the replication that follows it
- Written quote after one meeting The price is discussed in a meeting about your systems
Evaluating a data replication platform means scoring each candidate on the same criteria with the same test data. This guide gives a scorecard with nine weighted criteria, the evidence to request for each, and a four-step test plan that ends in a parallel run. It is written for data and procurement teams in Saudi Arabia.
The scorecard
The weights are a suggestion. Change them to suit your project, keep the total at 100, and print one card for each vendor.
| Criterion | Suggested weight | Evidence to request | Score (1 to 5) |
|---|---|---|---|
| Who makes it | 10 | Where the vendor is based and who represents it in Saudi Arabia, in writing | 1 · 2 · 3 · 4 · 5 |
| Licensing model | 10 | The licensing model and the quote, both in writing | 1 · 2 · 3 · 4 · 5 |
| Who runs the tool | 15 | Who operates it day to day, your team or the vendor's, and which regular reports you receive | 1 · 2 · 3 · 4 · 5 |
| Local support | 10 | The support language and working hours, written into the contract | 1 · 2 · 3 · 4 · 5 |
| Where it runs | 10 | Where the platform runs, and whether any data leaves the Kingdom | 1 · 2 · 3 · 4 · 5 |
| Sources and targets | 15 | The supported sources and targets with their versions, checked against yours | 1 · 2 · 3 · 4 · 5 |
| Custom connectors | 5 | What happens when a system is not on the list, and who builds the connector | 1 · 2 · 3 · 4 · 5 |
| Two-way synchronisation | 10 | Whether both sides can take writes, and how a conflict is resolved | 1 · 2 · 3 · 4 · 5 |
| Load control on the source | 15 | How transfer speed and resource use are limited on production | 1 · 2 · 3 · 4 · 5 |
Multiply each score by its weight, add the nine results and divide by five: the total is out of 100. For the questions behind each criterion and how iblync answers them, see nine criteria for comparing data replication tools.
Set the weights before you meet a vendor
The weights decide the result more than the scores do, so fix them before the first vendor presentation. A team that has just seen a demonstration tends to raise the weight of whatever that vendor showed best.
Agree the weights in one meeting with the owners of the source and target databases, security and procurement. Record them, and do not change them once vendor meetings begin. If your project has a condition it cannot give up, such as data staying in the Kingdom or a source version that must be supported, write it above the card as a pass or fail condition rather than as a heavy weight: a vendor that fails it is not scored at all.
Evidence to request from every vendor
Ask for each item in writing. A spoken answer in a meeting is not evidence.
- The list of supported sources and targets, with versions, checked against your exact systems.
- The licensing model in writing, together with the written quote.
- Who operates the platform day to day and who supports it, in which language and in which hours.
- Where the platform runs, and whether any data leaves the Kingdom at any stage.
- What the platform does with changes when the target is unavailable, and how it catches up when the target returns.
- How the load on the production database is limited: transfer speed and resource use.
- An exit plan: how you would stop using the platform, and what stays with you.
- A reference you can call: a team that runs the platform today.
A four-step test plan
Run the same plan, with the same data, on every platform on the shortlist. The plan sets no durations: each step ends when its pass condition is met.
-
One set of test data.
Prepare one copy of data at real size, with your heaviest tables, your deletes and the data types that cause trouble, and give the same copy to every vendor.
-
Initial load.
Load the data once into the target with the load on the source limited, then check row counts and sample values against the source.
-
Continuous sync, compared.
Keep writing to the source, including updates and deletes, and compare the two sides. Stop the target for a while and check that no change is lost when it returns.
-
Parallel run on a limited scope.
Run the old and the new side by side for one application or one group of tables, and judge the result against the weights you set. Our page on zero downtime migration explains how parallel running and the switch work.
Who should be in the room
Four roles score together, and each owns part of the card. The owner of the source database judges the load on production and the version fit. The owner of the target judges what arrives and how it is checked. Security judges where the platform runs, who can reach the data and what stays inside your environment. Procurement keeps the weights fixed, collects the written evidence and makes sure every vendor answered the same questions.
A score that one of them gave is an opinion. A score all four agreed is a decision that stands up in procurement review.
Five mistakes in replication evaluations
- Testing with small data. A few thousand rows load cleanly on any platform. Test with your largest tables and a realistic rate of change.
- Ignoring deletes. Many tests check inserts and updates only. Delete rows on the source and confirm they are gone from the target.
- Ignoring a stopped target. Stop the target during the test and see whether changes wait in a queue or are lost.
- Comparing quotes without comparing operation. Two quotes say little until you know who runs each platform after go-live, and how.
- Forgetting the exit plan. Ask how you would leave the platform before you sign, not after.
How iblync is evaluated
Score iblync on the same card as any other platform. It is built and supported by Interkey, a Saudi technology company in Riyadh, and delivered with the Interkey team that operates it, with no development team needed on your side, and with weekly and monthly replication reports. Support comes from Interkey, in Arabic and English. Where it runs is decided per project, before design. Custom connectors are built when needed. Two-way synchronisation is supported, transfer speed and resource use can be limited, and changes wait in a queue while the target is stopped.
The price is discussed in a meeting, followed by a written quote. Read more on the iblync platform, its connectors and how pricing works.
Questions teams ask when evaluating replication platforms
How do you evaluate a data replication platform?
With fixed weighted criteria, written evidence for each criterion, and a test on your own data that ends in a parallel run. Set the weights before you meet any vendor, give every vendor the same test data, and score with the owners of the source and target, security and procurement together.
What should a data replication RFP include?
Your sources and targets with their versions, the data volume and how fast it changes, the downtime window the business will accept, where the platform may run, and the support you need: language, hours and who operates the platform. Ask every vendor to answer in writing, against the evidence list on this page.
How long should the test of a replication platform run?
Until it has covered a full business cycle with your heaviest operations, such as month-end processing or a peak season. A fixed number of days means little: a test that misses your heaviest load has not tested what will matter in production.
What is the most overlooked criterion?
Who runs the platform after purchase, and what happens to changes when the target stops. Both are easy to skip in a demonstration and costly to discover later. Ask every vendor for a written answer to both, and test the second by stopping the target during your test.
Can we use the scorecard as it is?
Yes. The scorecard is the table on this page: print it or save it as a PDF with the button at the top, one copy for each vendor. Change the weights to suit your project before scoring, and score every vendor on the same criteria and the same test data.
Next
Read next
Next step
Ask an engineer one question
Tell us the sources and targets you are evaluating and what you need to decide. Ask one question about your evaluation.
What happens next
- We read your request
- An engineer contacts you
- You get an estimate or a demo
Or directly
+966-11-2180999 [email protected]Tawuniya Towers, North Tower, 7th Floor, King Fahad Highway, Olaya, P.O. Box 56835, Riyadh 11564, Saudi Arabia