欠陥のあるマザーボードコントローラを悪用して数千台のサーバにバックドアが仕掛けられる恐れThousands of servers can be backdoored by exploiting buggy motherboard controllers
匿名の公開いいねです。記事の保存・お気に入りではなく、Featured、Top 3、重要度、掲載順位には影響しません。仕組みとプライバシーAnonymous public likes are reactions, not saved articles or bookmarks. They do not affect Featured, Top 3, importance, or listing order.How it works and privacy
- 大手メーカー製のベースボード管理コントローラ(BMC)に深刻な脆弱性が発見され、攻撃者がサーバにバックドアを設置できる状態にあることが明らかになった。
- BMCはOS外で動作するため、侵害されると検出・除去が極めて困難になる。
- Serious security flaws in baseboard management controllers from major manufacturers leave thousands of servers open to persistent backdoors.
- Because BMCs operate below the OS layer, successful exploitation is extremely difficult to detect or remediate.
要約と収集メタデータをもとに生成した AI 解説本文です。元記事全文の転載・翻訳ではありません。This AI explainer is generated from the summaries and collected metadata, not from a reproduction or translation of the full source article.
サーバのマザーボードに搭載される管理チップ「ベースボード管理コントローラ(BMC)」に深刻な脆弱性が見つかり、攻撃者が数千台規模のサーバに検出困難なバックドアを仕掛けられる恐れがあると報じられた。大手メーカー製の製品が広く影響を受けるとされ、企業のデータセンターやクラウド基盤にとって看過できない問題となっている。
BMCは、サーバ本体のOSやCPUとは独立して動作する小型の管理用プロセッサだ。管理者はこれを通じて、電源のオン・オフや遠隔での再起動、ハードウェアの温度・電圧監視、ファームウェアの更新といった作業を、サーバが停止している状態でもネットワーク越しに行える。DellのiDRAC、HPEのiLO、あるいは業界標準のIPMIなどが代表的な仕組みで、大規模なサーバ運用には欠かせない存在となっている。
問題の深刻さは、BMCが動作する「層」にある。BMCはOSよりも下位で稼働するため、いったん侵害されると、OS上で動くウイルス対策ソフトやログ監視ツールからはその挙動がほとんど見えない。攻撃者が制御を奪えば、OSを再インストールしても、あるいはストレージを入れ替えても、バックドアが残り続ける可能性がある。検出も除去も極めて難しく、永続的な足がかりを与えかねない点が懸念される。
大手メーカー製のベースボード管理コントローラ(BMC)に深刻な脆弱性が発見され、攻撃者がサーバにバックドアを設置できる状態にあることが明らかになった。
BMCをめぐるセキュリティ上の懸念は、以前から繰り返し指摘されてきた経緯がある。初期パスワードの使い回しや、管理インターフェースがインターネットに露出しているケースがたびたび報告されており、専門家は外部から直接アクセスできない隔離ネットワークでの運用を推奨してきた。ただし今回のように、コントローラやファームウェアそのものに欠陥が及ぶ場合、利用者側の設定だけでは防ぎきれない可能性がある。
影響を受ける組織には、メーカーが提供するファームウェア更新の有無を確認し、修正が公開され次第すみやかに適用する対応が求められると見られる。ハードウェアに近い層を狙う攻撃はサプライチェーン全体に関わるだけに、今後もメーカーと利用者双方の姿勢が問われることになりそうだ。
Security researchers have disclosed a set of serious vulnerabilities in the baseboard management controllers (BMCs) embedded in servers from several of the world's largest hardware manufacturers, warning that the flaws could allow attackers to plant persistent backdoors on thousands of machines. The disclosure matters because BMCs operate beneath the operating system, giving an intruder a foothold that is exceptionally difficult to detect and even harder to remove once established.
A BMC is a small, independent computer soldered onto a server motherboard. It has its own processor, memory, firmware, and often a dedicated network connection, and it runs whether or not the host system is powered on. Administrators rely on this out-of-band access to monitor hardware health, reinstall operating systems, mount virtual media, and reboot or reconfigure machines remotely—capabilities that are indispensable in large data centers where physically touching each server is impractical. The same deep privileges that make BMCs useful also make them a high-value target.
Because a BMC sits below the OS layer, code running on it is effectively invisible to the endpoint detection tools, antivirus software, and logging agents that defenders typically deploy inside the operating system. An attacker who gains control of the controller can potentially survive OS reinstalls, disk wipes, and in some cases even firmware updates, while reaching into the host to read memory, alter boot processes, or re-implant malware. That combination of stealth and durability is why the exposure is being described as particularly dangerous, and why researchers characterize the current state of BMC security as a mess.
BMC functionality is commonly exposed through the Intelligent Platform Management Interface (IPMI), an older standard long criticized for weak authentication, and increasingly through Redfish, a more modern REST API defined by the DMTF standards body. Major vendors layer their own management stacks on top of these controllers, such as Dell's iDRAC, Hewlett Packard Enterprise's iLO, Lenovo's XClarity, and Supermicro's implementations. Much of the underlying firmware across the industry originates from a small number of suppliers—American Megatrends' MegaRAC being one of the most widely deployed—which means a single defect can ripple across many otherwise unrelated brands.
This is not the first time BMCs have drawn scrutiny. In 2022 and 2023, firmware security firm Eclypsium published a series of flaws in MegaRAC, collectively branded "BMC&C," that similarly enabled remote code execution and persistence. The recurrence of comparable issues suggests that supply-chain concentration and the difficulty of patching low-level firmware remain unresolved structural problems rather than isolated bugs. When many server brands depend on shared components, a weakness in that common code can propagate widely before it is understood or fixed.
Serious security flaws in baseboard management controllers from major manufacturers leave thousands of servers open to persistent backdoors.
The practical risk depends heavily on exposure. Management interfaces that are reachable from the public internet, or that share a network with general workloads, dramatically widen the attack surface. Security guidance for BMCs has been consistent for years and appears to apply here as well: keep management interfaces off the public internet, place them on isolated management networks, disable unused or legacy protocols such as older IPMI variants, enforce strong and unique credentials, and apply vendor firmware updates promptly. Organizations that cannot immediately patch are likely to rely on network segmentation to limit exposure while fixes are tested and rolled out. Firmware updates on controllers of this kind can be slow and operationally sensitive, since a failed flash can disable remote management or, in the worst case, the server itself.
For enterprise and cloud operators, the concern is scale. A data center may contain thousands of identical servers, and a flaw affecting a common controller could, in principle, touch a large fraction of a fleet at once. A compromised BMC is also an attractive pivot point for lateral movement, because it can influence the host it manages and may hold administrative credentials.
The full scope of the issue—including exactly which models are affected, whether any of the flaws have been exploited in the wild, and the timeline for coordinated patches—will likely become clearer as vendors publish advisories and remediation guidance. In the meantime, the disclosure is a reminder that the trusted, below-the-OS components that keep modern servers manageable can themselves become the weakest link, and that securing infrastructure increasingly means looking beneath the software stack most defenses are built to watch.
本ページの本文と要約は AI による自動生成です。日本語版と英語版は言語ごとに独立して生成されるため、表現や詳しさが異なる場合があります。正確性は元記事 (arstechnica.com) をご確認ください。The body and summaries are AI-generated independently for each language, so wording and detail may differ. Verify accuracy at the original source (arstechnica.com).





