To cite an archived website in IEEE, identify the original author and page, verify the particular snapshot you read, and use its complete archived URL. Keep the archive's capture date separate from the source's publication date and your access date. A saved error page cannot support the missing content.
We checked a former Google Labs page, its readable Wayback capture, and IEEE's website reference pattern. The HTTPS form of the old MapReduce address returned 404 on October 30, 2026. Its June 12, 2006 capture still displayed the title, two credited authors, and an abstract. The restored page lets us check its identity and read the evidence instead of guessing from the broken link.
Identify the original page
Start with the exact address in your notes or reference list. Save its path and query parameters, along with any title or author information recorded when you first used it. A domain-only search can lead to a homepage while the paper sits several folders deeper.
Our historical address is http://labs.google.com/papers/mapreduce.html. The archived heading identifies MapReduce and credits Jeffrey Dean and Sanjay Ghemawat. The page sits under Research Publications and links to a conference paper and slides. These details distinguish this landing page from an unrelated page mentioning the same method.
Record what the current address actually does. A 404, a redirect to a new location, and a page whose text has changed are different observations. Follow a redirect and compare the destination with your notes; the publisher may have moved the material intact. A failure today does not tell you the date when the original vanished.
Decide which object supports your sentence. This example cites the archived landing page as evidence of what that page displayed. A claim drawn from the complete research paper should cite the paper. Our conference-paper guide explains the proceedings fields. Restoring a link does not change the publication type of the source behind it.
Find the relevant capture
This site's generator does not restore missing pages. Enter the full original address in the Wayback Machine and inspect its calendar. Choose a capture relevant to the period you are discussing or the version recorded in your notes. A convenient recent capture may preserve a redirect or an error instead of the old content.
For this example, the calendar led to the readable address below.
Open the June 12, 2006 capture of the Google Labs page.
The fourteen digits identify the saved capture. Here they encode June 12, 2006, at 23:02:33. Copy the URL after the page has finished loading. A requested date may resolve to a different available capture, so the final address is the one to record. A calendar search URL with an asterisk does not identify one version.
The capture shows a page that existed in the archive at that time. It does not establish the page's first publication date. Nor does choosing an earlier capture prove that every sentence remained unchanged throughout the years between that capture and your original visit.
Verify that the evidence survives
Read the paragraph that supports your claim. Check the heading and author credit, then inspect the surrounding context for qualifications. The recovered MapReduce page contains three abstract paragraphs and an Appeared in entry naming OSDI'04. Its Google Labs logo also loaded in our check. We did not infer completeness from a successful page load alone.
Linked material needs another check. The page's PDF and slide links are separate resources. The landing page can survive even when a linked document is missing. Follow the item you need and inspect its final archived address and visible content.
Consider an explicitly hypothetical incomplete capture. Suppose the heading and the first abstract paragraph survived, but the paragraph reporting usage figures was missing. You could identify the paper and discuss the preserved introductory description. You could not use that capture to verify the omitted figures by borrowing them from a newer page.
The same limit applies to an image. If a caption survives but its chart is absent, record the missing chart. Do not reconstruct bar heights or numerical values from the caption. Internet Archive explains that images can be missing and that features depending on the original server may fail.
Keep publication, capture, and access dates separate
The date labels on the MapReduce page need to be read in context. Its conference entry says December 2004, its footer says copyright 2005, and its Wayback address contains a 2006 capture date. None of those labels states when this webpage itself was first published or last updated.
| Date | Verified value | Meaning for the reference |
|---|---|---|
| Original publication or update | Paper publication shown as December 2004; webpage publication/update unstated | Retain the paper date for a paper reference. Do not assign it automatically to the webpage. |
| Archive capture | June 12, 2006, encoded by 20060612230233 | Identifies the archived version opened here. |
| Access | October 30, 2026 | Records when we consulted this capture. |
IEEE's basic website pattern includes the credited author, page title, website title, access date, and URL. It does not require inserting the capture date as a publication year. The worked entry below applies that general website pattern to the archived landing page. IEEE's guide does not give a separate Wayback Machine template in the sections consulted.
[1] J. Dean and S. Ghemawat. “MapReduce: Simplified Data Processing on Large Clusters.” Google Research Publications. Accessed: Oct. 30, 2026. [Online]. Available: https://web.archive.org/web/20060612230233/http://labs.google.com/papers/mapreduce.htmlKeep the original address and capture timestamp in your source record alongside the reference. If the destination requires both URLs or an explicit archive note, follow its instructions. Internet Archive's help page includes an MLA example; its punctuation should not be copied into an IEEE reference.
For ordinary missing-author or missing-date questions, use our website citation guide. For a versioned manual, the software-documentation guide explains how to retain the release information.