Playbook

How to modernize classic ASP website

A step by step playbook for how to modernize classic ASP website safely, page by page, without a big bang rewrite or a frozen roadmap.

By the Cloudverg engineering team · · 6 min read

The problem

How to modernize classic ASP website is usually framed as a choice between leaving it alone and rewriting everything. Both options are bad. The old site still takes orders, pays staff or serves customers. A rewrite takes months, and nobody can list every rule hidden in .asp pages, include files and COM components. The real problem is not VBScript. It is that the system has no map.

Classic ASP sites tend to carry the same weight. Pages mix HTML, SQL and business logic. Includes get copied between folders. Session state lives in the worker process, so a recycle logs people out. A 32-bit COM DLL needs a 32-bit app pool and registration on every server. Database drivers and TLS settings age out. The first failures are often small: a payment provider changes an endpoint, a mail relay stops accepting old cipher suites, or a report control will not install on a newer Windows Server.

A big-bang rewrite also freezes the business. Feature requests wait. Security fixes get applied to a codebase that is about to be thrown away. The team loses the people who know how the site really works. The safer path is to keep the old site running while you move it in slices, with a rollback at every step.

What a good solution looks like

A good plan treats the classic ASP site as a production asset, not a museum piece. Put a reverse proxy or load balancer in front of it early. That gives you one place for TLS, logging, redirects and routing. You can send /new to a new app and keep everything else on the old one. IIS URL Rewrite and Application Request Routing can do this on Windows. A small ASP.NET Core app or nginx can do it too.

Then decide per page, not per site. Each route gets one of four labels: keep, refactor, replace, retire. A contact form may be replaced in a day. A pricing engine may stay in classic ASP for a year. An admin report that nobody uses gets retired. Write these labels down with an owner and a rollback point. A good solution also keeps URLs alive: old .asp links redirect to new routes with a 301, and query strings that matter are mapped, not dropped.

Expect a hybrid for a while. Classic ASP can run on current Windows Server and IIS. Keep the 32-bit app pool if COM needs it. Move static files and images to the new app first. Centralize authentication with a signed cookie or an identity provider, then let both stacks read it. Use database views or stored procedures as a contract, so new code does not depend on table layouts that the old code still changes.

How to build it

Start with a baseline you can test. Crawl the site and export IIS logs for a month. List every page with traffic, every form post, every scheduled job and every external service call. Note which pages use COM, which use session state, and which write to the database. Build a smoke test list of the ten to twenty paths that must work. Check status codes, page titles and a few text markers with PowerShell, curl or a browser test tool. Save a copy of the site and database that you can restore.

Stand up the new environment beside the old one. Put the proxy in front. Route a single low-risk path, such as the privacy page or a read-only product list, to the new app. Keep the old path reachable at /legacy so support can compare. Add error logging and request tracing before you move real traffic. On the classic side, turn on detailed errors in a test environment, but keep custom errors on in production.

Move in this order: static assets, shared layout, authentication, read-only pages, forms, then admin. For auth, issue a cookie at the proxy or identity provider and validate it in both stacks. Do not try to share classic ASP session state directly. Classic ASP keeps session data in the IIS worker process. Store the shared values in a database table or a cache with your own token. For each form, capture the exact field names and validation rules before you rewrite it. Test file uploads, email sending and redirects after the move.

For COM components, choose one approach per component. If it works and is stable, keep it in a 32-bit app pool behind a small service. If it wraps a simple rule, rewrite it in the new stack with tests. If it talks to hardware or a vendor system, wrap it and log every call. Never move the database and the code on the same day. Redirect old .asp URLs one by one, keep a mapping file, and retire a page only when logs show no traffic for a full business cycle.

Risks and how to manage them

The biggest risk is hidden logic. Include files, global.asa and copy-pasted functions often hold rules that nobody wrote down. Read them, then interview the people who use the site. Ask what they do when a page fails. That answer often reveals a manual workaround that the code never handled. Keep a risk list ranked by outage cost and likelihood, not by how interesting the code is.

Session and auth breaks are common. A user logs in on the old page, clicks a new link, and lands logged out. Test the full path with a real account before each release. Watch cookie names, domains, paths, SameSite settings and clock skew. Encoding is another quiet failure: classic ASP often writes Windows-1252, while a new app defaults to UTF-8. Check names, addresses and currency symbols in forms and reports.

Operational risks matter too. The app pool identity may have database rights that the new app does not. Scheduled tasks may run under a personal account. Backups may exist but never be restored. Keep the old server running until the new one has handled a full billing or order cycle. Have a rollback that takes minutes: flip the proxy route back, restore the last database backup if needed, and tell support which URL to use.

How to measure success

Measure the migration, not the technology. Track the count of pages moved, the count of old URLs still receiving traffic, error rate by route, login success, form submission success and job completion. Add response time for the critical paths. Keep a simple dashboard that support and engineering both read. If a page move causes a spike in support tickets, stop and fix it before moving the next page.

Deployment health matters as much as page health. Record how long a rollback takes, how often releases happen, and how many releases need a hotfix. A good sign is that the old site becomes boring: fewer errors, fewer manual restarts, no surprise outages. Another sign is that new features land in the new stack without touching the classic code.

When the old site has no traffic and no scheduled jobs, turn it off. Keep the code and database backup for the retention period your business requires. Then remove the proxy route and close the firewall rules. If you want a fixed scope start, Cloudverg's Legacy modernization sprint is $8,500/month for a fixed scope of three weeks. It inventories the apps, recovers business rules, ranks risks and gives you a step by step migration sequence with rollback points. You can also use an internal architect or a freelance assessor. The point is the same: get the map before you move the walls.

Frequently asked questions

Can we keep classic ASP and still modernize the site?

Yes. Run classic ASP on current Windows Server and IIS, put a proxy in front, and move pages one at a time. Keep 32-bit app pools for COM components. The old and new stacks can share a login cookie and a database contract while you migrate.

Should we rewrite everything at once?

No. A big-bang rewrite hides unknown rules and freezes feature work. Move static assets and auth first, then read-only pages, forms and admin. Keep a rollback route at the proxy so you can switch back in minutes.

How do we handle classic ASP session state during migration?

Do not try to share the in-process session directly. Issue your own signed token at the proxy or identity provider and store shared values in a database or cache. Validate the token in both stacks. Test login, logout and timeout paths with real accounts.

Monthly plan

Want help with this? Legacy modernization sprint

$8,500/month

3 weeks · Starts within 10 business days

Three weeks to understand an old system nobody wants to touch, rank the risks and agree a step by step plan that avoids a big bang rewrite.

Renews monthly. Cancel any time.
What is included

Related guides

More custom software guides   Custom software

Smallest way to start

Get a senior architect on your problem this week

Ninety minutes a month on the cloud, AI or security question that is blocking you, a pre-read of your material and a written summary within two business days.

$450/month

Renews monthly. Cancel any time.