<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Hans Study</title>
    <description>The latest articles on DEV Community by Hans Study (@hansstudy).</description>
    <link>https://dev.to/hansstudy</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3949218%2F60c94993-db08-4321-b379-507b50d567ea.jpeg</url>
      <title>DEV Community: Hans Study</title>
      <link>https://dev.to/hansstudy</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/hansstudy"/>
    <language>en</language>
    <item>
      <title>Vendor Remote Access to OT Without a Forgotten VPN</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Sat, 14 Nov 2026 19:23:41 +0000</pubDate>
      <link>https://dev.to/hansstudy/vendor-remote-access-to-ot-without-a-forgotten-vpn-3g1i</link>
      <guid>https://dev.to/hansstudy/vendor-remote-access-to-ot-without-a-forgotten-vpn-3g1i</guid>
      <description>&lt;p&gt;Give vendors access to a jump host in the industrial DMZ, not a route into the control network. Open it for a named person, a named target, and a time window. Require MFA to reach the jump host, record the session, and close it when the work ends.&lt;/p&gt;

&lt;p&gt;Most remote access problems I see are not exotic. A VPN account made for a commissioning job is still active 3 years later. A cellular router on a panel still has its default password. Nobody owns the list. CISA’s guidance on reducing OT cyber threats lists limiting exposure of OT to public networks and securing remote access among its primary mitigations, and NIST SP 800-82 Rev. 3 treats remote access as a boundary that needs its own controls.&lt;/p&gt;

&lt;h2&gt;
  
  
  The pattern
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; The vendor authenticates to a gateway with MFA.&lt;/li&gt;
&lt;li&gt; The gateway hands them a session on a jump host in the iDMZ (Level 3.5 in the &lt;a href="https://hans.study/purdue-model-in-practice-vs-on-paper/" rel="noopener noreferrer"&gt;Purdue model&lt;/a&gt;).&lt;/li&gt;
&lt;li&gt; From the jump host, a firewall rule allows one inbound path to one named target, such as the engineering workstation.&lt;/li&gt;
&lt;li&gt; The session is recorded, and the recording is kept.&lt;/li&gt;
&lt;li&gt; The access window closes by itself.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The vendor’s own laptop never has an IP route to a PLC. If the vendor needs their programming software, install it on the engineering workstation or a controlled virtual machine behind the jump host. This is the remote access conduit in the &lt;a href="https://hans.study/iec-62443-zones-conduits-small-plant-example/" rel="noopener noreferrer"&gt;zones and conduits worked example&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  The controls, and what each one stops
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Control&lt;/th&gt;
&lt;th&gt;What it stops&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Jump host in the DMZ&lt;/td&gt;
&lt;td&gt;A compromised vendor laptop reaching the plant directly&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Per-person accounts&lt;/td&gt;
&lt;td&gt;Shared logins nobody can attribute&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;MFA&lt;/td&gt;
&lt;td&gt;Stolen or reused passwords&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Time-bound access&lt;/td&gt;
&lt;td&gt;The forgotten account and the always-open tunnel&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Session recording&lt;/td&gt;
&lt;td&gt;Disputes over what changed, and slow incident review&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;One named target&lt;/td&gt;
&lt;td&gt;Lateral movement from the engineering host to everything else&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Approval by the operator&lt;/td&gt;
&lt;td&gt;Surprise sessions during production&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Time-bound access works best when someone at the plant approves each session by phone or ticket. It is a small step, and it catches the unexpected login, which is the one you most want to catch.&lt;/p&gt;

&lt;h2&gt;
  
  
  View-only first
&lt;/h2&gt;

&lt;p&gt;Many support calls are diagnostic. Some jump host and remote access products can give a view-only session first, with a second approval for a session that can change logic. Not every tool supports this, so ask.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to ask the vendor
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; Do you connect through our jump host, or do you require your own tunnel or modem?&lt;/li&gt;
&lt;li&gt; Are your technicians named, and do they use MFA?&lt;/li&gt;
&lt;li&gt; Which ports and protocols do you need, and to which devices?&lt;/li&gt;
&lt;li&gt; What remote support tools run on your laptops, and who patches them?&lt;/li&gt;
&lt;li&gt; Do you keep copies of our project files, and where are they stored?&lt;/li&gt;
&lt;li&gt; How do you tell us if you suspect your own environment was compromised?&lt;/li&gt;
&lt;li&gt; Do your practices align with IEC 62443-2-4 for service providers?&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;If a vendor insists on a permanent cellular modem inside a panel, record it as an exception with an owner and a review date. Do not let it go unrecorded.&lt;/p&gt;

&lt;h2&gt;
  
  
  When the vendor pushes back
&lt;/h2&gt;

&lt;p&gt;Some vendors want their own tunnel because it is how they support every customer. That is a reasonable thing to say and a poor reason to accept it. Offer the jump host, a test window, and a named contact on your side. A vendor that cannot work through a customer-controlled gateway is worth recording as a supplier risk, with an owner and a review date.&lt;/p&gt;

&lt;h2&gt;
  
  
  Hardening the jump host
&lt;/h2&gt;

&lt;p&gt;Treat it as a sacrificial, well-watched system. Patch it, keep it off the business domain if you can, disable clipboard and file transfer unless the job needs them, and send its logs somewhere the vendor cannot reach. Block outbound internet from it except what the gateway needs.&lt;/p&gt;

&lt;p&gt;Remote access also leaves a trail you can use. Pair it with &lt;a href="https://hans.study/ot-asset-inventory-without-active-scanning/" rel="noopener noreferrer"&gt;passive monitoring and an inventory&lt;/a&gt; so a session that touches an unexpected device shows up.&lt;/p&gt;

&lt;p&gt;The wider set of controls is in &lt;a href="https://hans.study/ot-network-security-controls/" rel="noopener noreferrer"&gt;Security controls for OT networks that hold up in production&lt;/a&gt;, and the site overview is on &lt;a href="https://hans.study/industrial-ot-networks/" rel="noopener noreferrer"&gt;industrial OT networks&lt;/a&gt;. The boundary design is taught in &lt;a href="https://hans.study/training/hardening-control-networks/" rel="noopener noreferrer"&gt;Hardening Control Networks&lt;/a&gt; and covered in the &lt;a href="https://hans.study/learning/ot-networks-level-1/" rel="noopener noreferrer"&gt;free OT networks course&lt;/a&gt;. The &lt;a href="https://hans.study/advisory/ot-it-convergence/" rel="noopener noreferrer"&gt;OT and IT convergence advisory&lt;/a&gt; covers the organizational side.&lt;/p&gt;

&lt;p&gt;If you want help designing or reviewing a vendor access path, see &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;/book/&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>networking</category>
      <category>security</category>
      <category>cybersecurity</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>The Purdue Model in Practice Versus on Paper</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Fri, 13 Nov 2026 20:57:32 +0000</pubDate>
      <link>https://dev.to/hansstudy/the-purdue-model-in-practice-versus-on-paper-1lea</link>
      <guid>https://dev.to/hansstudy/the-purdue-model-in-practice-versus-on-paper-1lea</guid>
      <description>&lt;p&gt;The Purdue model is a layered way to describe an industrial network, from the process at Level 0 up to the enterprise at Level 5. Real plants blur it. The fix that follows the model is an industrial DMZ at Level 3.5, with nothing on the business side opening a session into the plant.&lt;/p&gt;

&lt;p&gt;The levels, briefly. Level 0 is sensors and actuators. Level 1 is controllers such as PLCs and RTUs. Level 2 is supervision, meaning HMIs, SCADA, and engineering workstations. Level 3 is operations management, including historians, batch, and MES. Level 4 is site business systems, and Level 5 is the enterprise. The &lt;a href="https://hans.study/learning/ot-networks-level-1/" rel="noopener noreferrer"&gt;free OT course&lt;/a&gt; walks through each one.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where plants blur the levels
&lt;/h2&gt;

&lt;p&gt;Paper diagrams show clean layers. Sites show things like these.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt; One flat network carries Level 1 through Level 3, so the historian and the PLCs share a broadcast domain.&lt;/li&gt;
&lt;li&gt; The engineering workstation is also the vendor’s remote access point and somebody’s email machine.&lt;/li&gt;
&lt;li&gt; An MES or reporting server reads PLC tags directly from Level 3 and is also reachable from the business network.&lt;/li&gt;
&lt;li&gt; A cellular router or wireless link connects a remote station straight to a cloud portal, skipping every level.&lt;/li&gt;
&lt;li&gt; Smart drives, cameras, and meters on the plant network have web interfaces and run firmware nobody tracks.&lt;/li&gt;
&lt;li&gt; A single firewall holds the business and the plant on two interfaces with a broad allow rule between them.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;None of this is unusual, and it does not mean the model is wrong. It means the model describes intent, and the plant has a history.&lt;/p&gt;

&lt;h2&gt;
  
  
  What the Level 3.5 iDMZ does
&lt;/h2&gt;

&lt;p&gt;The industrial DMZ is a controlled network between Level 3 and Level 4. It was not in the original model. It is where the brokered services live.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Lives in the iDMZ&lt;/th&gt;
&lt;th&gt;Why&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Jump host or remote access gateway&lt;/td&gt;
&lt;td&gt;Outside sessions end here, not in the plant&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Historian replica&lt;/td&gt;
&lt;td&gt;The business reads a copy, not the plant source&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Patch and antivirus update servers&lt;/td&gt;
&lt;td&gt;Updates are staged and then pushed in&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;File transfer staging&lt;/td&gt;
&lt;td&gt;Files cross in a controlled, scanned way&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The rule is that no session crosses it end to end. The business side connects in, the plant side connects out, and the DMZ host terminates both. A compromise of the business network lands in the DMZ rather than at the controllers. For the access side of this, see &lt;a href="https://hans.study/vendor-remote-access-ot-without-forgotten-vpn/" rel="noopener noreferrer"&gt;Vendor remote access to OT without the forgotten VPN&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why the blur matters
&lt;/h2&gt;

&lt;p&gt;Each blurred level is a path a mistake or an attacker can use. A shared engineering workstation means one phished mailbox can reach the programming tools. A flat Level 1 to Level 3 network means a heavy historian query can load the same switch that carries controller traffic. Neither needs a sophisticated attacker, only an ordinary bad day. Naming the blur is how you decide which one to fix before the next outage window.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to fix first
&lt;/h2&gt;

&lt;p&gt;Order matters, because some fixes make the next one possible.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt; Inventory the devices, passively. See &lt;a href="https://hans.study/ot-asset-inventory-without-active-scanning/" rel="noopener noreferrer"&gt;Building an OT asset inventory without active scanning&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt; Find every path between the business network and the plant, including modems and vendor tunnels.&lt;/li&gt;
&lt;li&gt; Replace those paths with a DMZ. Even a small one beats a direct allow rule.&lt;/li&gt;
&lt;li&gt; Separate the engineering workstation from general use.&lt;/li&gt;
&lt;li&gt; Split Level 1 from Levels 2 and 3 with a firewall that knows the protocols in use.&lt;/li&gt;
&lt;li&gt; Add passive monitoring so you notice what changes.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Use zones and conduits to finish the job
&lt;/h2&gt;

&lt;p&gt;Purdue tells you the layers. IEC 62443-3-2 tells you how to group assets into zones and name the conduits between them, including those that do not fit a layer, such as the cellular link. A worked small-site version is in &lt;a href="https://hans.study/iec-62443-zones-conduits-small-plant-example/" rel="noopener noreferrer"&gt;IEC 62443 zones and conduits for a small plant&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Do not redraw the whole network on paper before changing anything. A plant that moves its vendor access and historian copy into a DMZ has improved more than one that has a perfect diagram and the same flat network.&lt;/p&gt;

&lt;p&gt;The broader controls are covered in &lt;a href="https://hans.study/ot-network-security-controls/" rel="noopener noreferrer"&gt;Security controls for OT networks that hold up in production&lt;/a&gt;. See also &lt;a href="https://hans.study/industrial-ot-networks/" rel="noopener noreferrer"&gt;industrial OT networks&lt;/a&gt;, the &lt;a href="https://hans.study/advisory/ot-it-convergence/" rel="noopener noreferrer"&gt;OT and IT convergence advisory&lt;/a&gt;, and the training course &lt;a href="https://hans.study/training/hardening-control-networks/" rel="noopener noreferrer"&gt;Hardening Control Networks&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you want your site reviewed against this list, see &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;/book/&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>networking</category>
      <category>security</category>
      <category>cybersecurity</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>OT Asset Inventory Without Active Scanning</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Thu, 12 Nov 2026 21:32:27 +0000</pubDate>
      <link>https://dev.to/hansstudy/ot-asset-inventory-without-active-scanning-1j3h</link>
      <guid>https://dev.to/hansstudy/ot-asset-inventory-without-active-scanning-1j3h</guid>
      <description>&lt;p&gt;You can build a usable OT inventory without sending a single packet at a controller. Combine passive traffic monitoring, switch MAC and port tables, the engineering project files, and a physical walk-down, then reconcile them in one list. Each source misses things the others catch.&lt;/p&gt;

&lt;p&gt;An inventory comes first because you cannot segment, patch, or monitor what you cannot name. The &lt;a href="https://hans.study/iec-62443-zones-conduits-small-plant-example/" rel="noopener noreferrer"&gt;zones and conduits worked example&lt;/a&gt; assumes you have one.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why active scanning is a risk
&lt;/h2&gt;

&lt;p&gt;An IT vulnerability scanner sends many probes to many ports quickly. Some older PLCs, drives, and embedded devices have small network stacks and may respond badly to unexpected packets. Effects range from a slowed scan cycle and dropped connections to a device fault. NIST SP 800-82 Rev. 3 notes that active scans may cause device instability or interfere with the device process state. The outcome depends on the device and firmware, so the safe default is not to test it on production.&lt;/p&gt;

&lt;p&gt;Active methods are not forbidden. They are something you choose, with a test unit, a window, and the vendor’s agreement.&lt;/p&gt;

&lt;h2&gt;
  
  
  Method 1, passive monitoring
&lt;/h2&gt;

&lt;p&gt;Connect a monitoring sensor to a SPAN port or a network tap on the switch that carries control traffic. It listens only. Over days it learns which addresses talk and which protocols they use, and often the vendor, model, and firmware from the protocol exchanges themselves.&lt;/p&gt;

&lt;p&gt;Known protocol ports help you recognize what you are seeing.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Protocol&lt;/th&gt;
&lt;th&gt;Port&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Modbus/TCP&lt;/td&gt;
&lt;td&gt;502/TCP&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;DNP3&lt;/td&gt;
&lt;td&gt;20000&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;EtherNet/IP&lt;/td&gt;
&lt;td&gt;44818/TCP, 2222/UDP&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Siemens S7&lt;/td&gt;
&lt;td&gt;102/TCP&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;OPC UA&lt;/td&gt;
&lt;td&gt;4840/TCP&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;BACnet/IP&lt;/td&gt;
&lt;td&gt;47808/UDP&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Place the sensor where the traffic passes. A sensor on the wrong switch sees nothing, and a single sensor at the top will not see traffic between two devices on the same access switch.&lt;/p&gt;

&lt;h2&gt;
  
  
  Method 2, switch tables
&lt;/h2&gt;

&lt;p&gt;Managed switches already know who is plugged in. The MAC address table, the ARP table on a router or layer 3 switch, and LLDP neighbour data give you addresses, port numbers, and sometimes device names. A read-only query to your own switch is far gentler than probing the end devices. Match the MAC prefix to a vendor and you have a first guess at what the device is.&lt;/p&gt;

&lt;h2&gt;
  
  
  Method 3, engineering project files
&lt;/h2&gt;

&lt;p&gt;PLC and HMI project files list I/O modules, network configuration, device IP addresses, and the other controllers each program talks to. They are some of the best inventory sources you have, and they cost nothing to read. Copy them offline and open them on the engineering workstation or a test machine. This also tells you whether a current backup exists, which is a finding on its own.&lt;/p&gt;

&lt;h2&gt;
  
  
  Method 4, the walk-down
&lt;/h2&gt;

&lt;p&gt;Walk the plant with a clipboard or tablet. Photograph nameplates, record serial numbers and firmware shown on the display, note unmanaged switches in cabinets, and look for the cellular router, the USB stick, and the laptop that was left plugged in. Passive tools do not see a device that is powered but silent, or a serial link with no Ethernet at all.&lt;/p&gt;

&lt;h2&gt;
  
  
  Reconcile into one list
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; Start with the project files, since they say what should exist.&lt;/li&gt;
&lt;li&gt; Add what passive monitoring and switch tables show.&lt;/li&gt;
&lt;li&gt; Walk down and resolve the differences.&lt;/li&gt;
&lt;li&gt; Flag unknown devices and ask the operators about them before you assume the worst.&lt;/li&gt;
&lt;li&gt; Assign an owner to every line.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The fields to record are in the FAQ above. Add the zone each device belongs to once the design exists. CISA’s asset inventory guidance and ISA/IEC 62443 both treat an up-to-date inventory as a basis for the rest of the program.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do with it
&lt;/h2&gt;

&lt;p&gt;Use it to find end-of-life devices, plan a patch or compensating control, spot paths into the plant such as &lt;a href="https://hans.study/vendor-remote-access-ot-without-forgotten-vpn/" rel="noopener noreferrer"&gt;vendor remote links&lt;/a&gt;, and compare the network against the layers in &lt;a href="https://hans.study/purdue-model-in-practice-vs-on-paper/" rel="noopener noreferrer"&gt;The Purdue model in practice versus on paper&lt;/a&gt;. It also gives you the baseline that &lt;a href="https://hans.study/ot-network-security-controls/" rel="noopener noreferrer"&gt;passive monitoring&lt;/a&gt; alerts against.&lt;/p&gt;

&lt;p&gt;More background is on &lt;a href="https://hans.study/industrial-ot-networks/" rel="noopener noreferrer"&gt;industrial OT networks&lt;/a&gt; and in the &lt;a href="https://hans.study/advisory/ot-it-convergence/" rel="noopener noreferrer"&gt;OT and IT convergence advisory&lt;/a&gt;. The monitoring lesson is in the &lt;a href="https://hans.study/learning/ot-networks-level-1/" rel="noopener noreferrer"&gt;free OT networks course&lt;/a&gt;, and the hands-on version is &lt;a href="https://hans.study/training/hardening-control-networks/" rel="noopener noreferrer"&gt;Hardening Control Networks&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you would like a hand building the first inventory, see &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;/book/&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>networking</category>
      <category>security</category>
      <category>cybersecurity</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>Door credentials for a defense supplier, and the held-open alarm</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Wed, 11 Nov 2026 21:31:01 +0000</pubDate>
      <link>https://dev.to/hansstudy/door-credentials-for-a-defense-supplier-and-the-held-open-alarm-1hmc</link>
      <guid>https://dev.to/hansstudy/door-credentials-for-a-defense-supplier-and-the-held-open-alarm-1hmc</guid>
      <description>&lt;p&gt;Most small defense suppliers have a lock on the server room door and a short list of people who hold a key. That list is the control, and it is also where the control fails, because keys leave no record. Nobody can tell you who opened the door last Tuesday.&lt;/p&gt;

&lt;p&gt;Physical access authorizations (03.10.01) and physical access control (03.10.07) are both in &lt;a href="https://hans.study/cpcsc/level-1/" rel="noopener noreferrer"&gt;CPCSC Level 1&lt;/a&gt;. An assessor wants a short list of who may enter, and some way to show that the list is what the door enforces.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why I push access control over keys
&lt;/h2&gt;

&lt;p&gt;Keys work. I recommend access control because it produces a record. Card or biometric beats a PIN, since a PIN can be known and a finger is harder to borrow. Card plus PIN belongs on the doors that matter, such as the server room and secure storage.&lt;/p&gt;

&lt;p&gt;The rest is simple and auditors like it. Employee ID on a navy lanyard, screened contractors on yellow, visitors on red, and a card design that shows division at a glance. Exit processes that pull card access the same day the person leaves.&lt;/p&gt;

&lt;h2&gt;
  
  
  The credential floor
&lt;/h2&gt;

&lt;p&gt;For credentials, I set the floor at 13.56 MHz Seos or iCLASS SE with custom keys, using HID's Elite program or equivalent, not the standard keyset. Use a corporate format such as HID Corporate 1000 or 2000.&lt;/p&gt;

&lt;p&gt;Custom keys matter because of what happened in February 2024. HID's advisories, including CVE-2024-23806, showed that keys could be pulled from iCLASS SE configuration cards and encoders. The standard keyset is the one everybody has, so custom keys are the floor.&lt;/p&gt;

&lt;p&gt;Three more settings follow from that:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Current firmware on the readers&lt;/strong&gt;, with configuration-card changes locked out, so a card in the wrong hands can't reconfigure a door.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Legacy and 125 kHz technologies disabled at the reader.&lt;/strong&gt; A reader that still accepts older technologies can be pushed down to them, so turn them off to stop the downgrade attack.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;OSDP Secure Channel between reader and controller.&lt;/strong&gt; It protects the link between the two. Readers and controllers sit on your network, so &lt;a href="https://hans.study/physical-security-network-hardening/" rel="noopener noreferrer"&gt;harden that network too&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Put all of it in the installer's scope of work, in writing, before the readers go on the wall.&lt;/p&gt;

&lt;h2&gt;
  
  
  Match it to what you make
&lt;/h2&gt;

&lt;p&gt;Match the controls to what you make. A washer supplier on a defense program is a defense contractor, and doesn't need a scramble pad. Anti-passback and tailgating controls scale from a sign to a turnstile to a two-door vestibule. The point is that the choice was made on purpose and can be explained.&lt;/p&gt;

&lt;h2&gt;
  
  
  Held-open alarms
&lt;/h2&gt;

&lt;p&gt;A held-open alarm fires when a door stays open longer than it should. On most access controllers it's a checkbox.&lt;/p&gt;

&lt;p&gt;It catches what the credential can't: the door that was opened legitimately and then left open. A propped door doesn't care how good your cards are, and the alarm is how you find out.&lt;/p&gt;

&lt;p&gt;Put held-open and forced-door alarms on the enclave doors, make sure someone receives them, and keep a record that they were tested. The record is the evidence. An alarm that nobody receives is a setting, not a control.&lt;/p&gt;

&lt;p&gt;The same goes for cameras. Cover the entries and main routes, and verify recording this month. Know the difference between a camera that shows someone is there and one that identifies a face. Thermal doesn't identify. Bispectral can. Put at least 1 identification-quality camera at the enclave entry.&lt;/p&gt;

</description>
      <category>security</category>
      <category>cybersecurity</category>
      <category>compliance</category>
      <category>networking</category>
    </item>
    <item>
      <title>IEC 62443 Zones and Conduits: Small Plant Example</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Wed, 11 Nov 2026 21:30:30 +0000</pubDate>
      <link>https://dev.to/hansstudy/iec-62443-zones-and-conduits-small-plant-example-1cdp</link>
      <guid>https://dev.to/hansstudy/iec-62443-zones-and-conduits-small-plant-example-1cdp</guid>
      <description>&lt;p&gt;A zone is a group of assets that share the same security requirements. A conduit is the controlled path between two zones, and it carries only what you have named. For a site with 3 PLCs, an HMI, a historian, an engineering workstation, and a vendor remote link, 5 zones and 5 conduits is a reasonable starting design.&lt;/p&gt;

&lt;p&gt;The standard behind this is IEC 62443-3-2, which covers risk assessment and the partitioning of a system into zones and conduits. IEC 62443-3-3 supplies the system requirements and security levels (SL) you check the result against. The vocabulary is covered in the &lt;a href="https://hans.study/learning/ot-networks-level-1/" rel="noopener noreferrer"&gt;free OT networks course&lt;/a&gt;, and the wider controls are in &lt;a href="https://hans.study/ot-network-security-controls/" rel="noopener noreferrer"&gt;Security controls for OT networks that hold up in production&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  The example site
&lt;/h2&gt;

&lt;p&gt;One building, one process line. 3 PLCs run the process, one HMI shows it to operators, a historian records tags, and an engineering workstation holds the PLC project files. A vendor logs in occasionally to support one PLC. The site has a business network next door for email and accounting.&lt;/p&gt;

&lt;h2&gt;
  
  
  The zones
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Zone&lt;/th&gt;
&lt;th&gt;Assets&lt;/th&gt;
&lt;th&gt;Why it is separate&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Control&lt;/td&gt;
&lt;td&gt;3 PLCs, their I/O&lt;/td&gt;
&lt;td&gt;Highest consequence if altered; fewest legitimate talkers&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Supervisory&lt;/td&gt;
&lt;td&gt;HMI&lt;/td&gt;
&lt;td&gt;Operators need it; it talks to the PLCs constantly&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Engineering&lt;/td&gt;
&lt;td&gt;Engineering workstation&lt;/td&gt;
&lt;td&gt;Can change PLC logic, so it needs the tightest access control&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Operations data&lt;/td&gt;
&lt;td&gt;Historian&lt;/td&gt;
&lt;td&gt;Receives data from the plant and is read by the business&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Remote access&lt;/td&gt;
&lt;td&gt;Jump host in an industrial DMZ&lt;/td&gt;
&lt;td&gt;The only place outside sessions terminate&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The business network sits outside these zones. It is untrusted from the plant’s point of view.&lt;/p&gt;

&lt;h2&gt;
  
  
  The conduits and their rules
&lt;/h2&gt;

&lt;p&gt;Write each rule as who, to what, over which port, and in which direction.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt; Supervisory to Control. The HMI may poll the PLCs on the protocol they use, for example EtherNet/IP on 44818/TCP and 2222/UDP, Modbus/TCP on 502, or S7 on 102. Nothing else.&lt;/li&gt;
&lt;li&gt; Engineering to Control. The engineering workstation may program the PLCs, and only when someone has switched it on. Keep it powered off or disconnected from the plant network when not in use if your process allows.&lt;/li&gt;
&lt;li&gt; Control to Operations data. The historian collects tags from the PLCs or from the HMI server. Prefer the historian initiating the connection, and read only.&lt;/li&gt;
&lt;li&gt; Operations data to business. Replicate historian data to a copy in the DMZ, and let the business read the copy. No business machine opens a session into the plant.&lt;/li&gt;
&lt;li&gt; Remote access to Engineering and Supervisory. The vendor connects to the jump host, and the jump host connects inward to one named target. The vendor’s laptop never gets a route to the PLC. How to build that is in &lt;a href="https://hans.study/vendor-remote-access-ot-without-forgotten-vpn/" rel="noopener noreferrer"&gt;Vendor remote access to OT without the forgotten VPN&lt;/a&gt;.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Every conduit ends with a deny rule that logs. The log is how you find the device nobody told you about.&lt;/p&gt;

&lt;h2&gt;
  
  
  What the security level means here
&lt;/h2&gt;

&lt;p&gt;IEC 62443 assigns each zone a target security level from SL 1 to SL 4, based on the kind of attacker you are defending against. SL 1 is protection against casual or coincidental violation. SL 2 is intentional violation using simple means and low resources. A small plant often lands on SL 2 for the control and engineering zones. That is a judgement made in the 3-2 risk assessment, and it should be written down with the reasoning. Component requirements live in 62443-4-2, and the supplier side is in 62443-2-4.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to build first
&lt;/h2&gt;

&lt;p&gt;If the plant is flat today, do not redraw everything in a weekend.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt; Put the vendor link behind a jump host.&lt;/li&gt;
&lt;li&gt; Move the historian copy into a DMZ so the business stops reaching the plant.&lt;/li&gt;
&lt;li&gt; Pull the engineering workstation into its own zone.&lt;/li&gt;
&lt;li&gt; Then split the HMI from the PLCs with a firewall that understands the protocols in use.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;An &lt;a href="https://hans.study/ot-asset-inventory-without-active-scanning/" rel="noopener noreferrer"&gt;asset inventory&lt;/a&gt; comes before any of it, because you cannot draw a zone around a device you do not know about. The model these zones map onto is discussed in &lt;a href="https://hans.study/purdue-model-in-practice-vs-on-paper/" rel="noopener noreferrer"&gt;The Purdue model in practice versus on paper&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where this goes wrong
&lt;/h2&gt;

&lt;p&gt;Zones drawn on a slide but not enforced by a firewall or managed switch are labels. Conduits with “any” in the port column are the same. And a flat VLAN with a firewall rule that says “OT to OT allow” has the zones in name only.&lt;/p&gt;

&lt;p&gt;For the broader picture see &lt;a href="https://hans.study/industrial-ot-networks/" rel="noopener noreferrer"&gt;industrial OT network security&lt;/a&gt; and the &lt;a href="https://hans.study/advisory/ot-it-convergence/" rel="noopener noreferrer"&gt;OT and IT convergence advisory&lt;/a&gt;. A hands-on version of this exercise is taught in &lt;a href="https://hans.study/training/hardening-control-networks/" rel="noopener noreferrer"&gt;Hardening Control Networks&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Want a second pair of eyes on your zone design? See &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;/book/&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>networking</category>
      <category>security</category>
      <category>cybersecurity</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>Cisco Catalyst IOS-XE Hardening: A Practical Baseline</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Tue, 10 Nov 2026 20:57:43 +0000</pubDate>
      <link>https://dev.to/hansstudy/cisco-catalyst-ios-xe-hardening-a-practical-baseline-19f9</link>
      <guid>https://dev.to/hansstudy/cisco-catalyst-ios-xe-hardening-a-practical-baseline-19f9</guid>
      <description>&lt;p&gt;A hardened Catalyst baseline means SSH version 2 only, central AAA with a local fallback, type 9 secrets, SNMPv3, authenticated NTP, and syslog off the box. On the access ports you add port security, BPDU guard, DHCP snooping, and Dynamic ARP Inspection, and you park every unused port in a blackhole VLAN. The one trap on camera networks is DAI blocking static-IP devices, which is covered below. Read that part first.&lt;/p&gt;

&lt;p&gt;The examples assume a Catalyst 9200 or 9300 on a current IOS-XE 17.x release, VLAN 99 for management, VLANs 100 to 102 for endpoints, and documentation addresses from 192.0.2.0/24. Replace every placeholder before use. The same ground is taught in &lt;a href="https://hans.study/learning/network-hardening-level-1/" rel="noopener noreferrer"&gt;Network Hardening Level 1&lt;/a&gt; and &lt;a href="https://hans.study/learning/cisco-catalyst-level-1/" rel="noopener noreferrer"&gt;Cisco Catalyst Level 1&lt;/a&gt;, and the audit side is in the &lt;a href="https://hans.study/standards-guidance/switch-configuration-audit-checklist-for-security-networks/" rel="noopener noreferrer"&gt;switch configuration audit checklist&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Management plane
&lt;/h2&gt;

&lt;p&gt;Start with identity. Create a local fallback account and a type 9 (scrypt) enable secret, then point logins at TACACS+ with local as the last method. Test the fallback from the console before you close your session.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;aaa new-model
&lt;span class="nb"&gt;enable&lt;/span&gt; algorithm-type scrypt secret &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt;
username breakglass privilege 15 algorithm-type scrypt secret &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt;
tacacs server TAC1
 address ipv4 192.0.2.20
 key &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt;
aaa group server tacacs+ TAC-GRP
 server name TAC1
aaa authentication login default group TAC-GRP &lt;span class="kd"&gt;local&lt;/span&gt;
aaa authorization &lt;span class="nb"&gt;exec&lt;/span&gt; default group TAC-GRP &lt;span class="kd"&gt;local&lt;/span&gt;
aaa accounting commands 15 default start-stop group TAC-GRP
login block-for 120 attempts 5 within 60
banner login ^C
This system is private property of &lt;span class="o"&gt;[&lt;/span&gt;ORGANIZATION&lt;span class="o"&gt;]&lt;/span&gt;. Authorized use only.
Activity is logged and monitored.
^C
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;SSH needs a domain name and a host key before it will run. Set version 2, restrict the VTY lines to the management subnet with an access class, and turn off everything else that listens.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;ip domain name example.net
crypto key generate rsa modulus 3072
ip ssh version 2
ip ssh time-out 60
ip ssh authentication-retries 3
ip access-list standard MGMT-ONLY
 permit 192.0.2.0 0.0.0.255
line vty 0 15
 access-class MGMT-ONLY &lt;span class="k"&gt;in&lt;/span&gt;
 transport input ssh
 exec-timeout 10 0
no ip http server
no ip http secure-server
no vstack
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The &lt;code&gt;no vstack&lt;/code&gt; line disables Smart Install, which has been exploited in the field and has no place on a production access switch. If you need the HTTPS interface, keep &lt;code&gt;ip http secure-server&lt;/code&gt; with an access class and a real certificate rather than the factory one.&lt;/p&gt;

&lt;h2&gt;
  
  
  Time, SNMP, and logging
&lt;/h2&gt;

&lt;p&gt;Logs without trustworthy time are hard to use. Incident reviews depend on them. Authenticate NTP so the switch only accepts time from a source it trusts, move to SNMPv3 with authentication and privacy, and send syslog to a collector that is not on the same switch stack. If your release does not offer a SHA-2 NTP key type, check its notes before you fall back to MD5.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;ntp authentication-key 1 hmac-sha2-256 &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt;
ntp authenticate
ntp trusted-key 1
ntp server 192.0.2.30 key 1
snmp-server group MON-GRP v3 priv
snmp-server user monitor MON-GRP v3 auth sha &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt; priv aes 128 &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt;
no snmp-server community public
service timestamps log datetime msec localtime show-timezone
logging source-interface Vlan99
logging host 192.0.2.10
logging &lt;span class="nb"&gt;trap&lt;/span&gt; informational
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Control plane basics
&lt;/h2&gt;

&lt;p&gt;Catalyst 9000 switches ship with a default control plane policing policy. Confirm it is attached with &lt;code&gt;show policy-map control-plane&lt;/code&gt;, and leave it alone unless you have measured a reason to change it, because a wrong policy can starve routing and management traffic during a burst. Measure first. Then remove the small things that widen the attack surface.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;no ip source-route
no ip domain lookup
no service pad
spanning-tree mode rapid-pvst
spanning-tree portfast bpduguard default
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Access ports
&lt;/h2&gt;

&lt;p&gt;Every port that faces an endpoint gets the same set of controls. Port security with a maximum of 1 MAC catches the unmanaged switch plugged in for extra drops, and BPDU guard catches a managed switch or a loop. Neither replaces the other.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;interface range GigabitEthernet1/0/1 - 23
 switchport mode access
 switchport access vlan 100
 switchport nonegotiate
 switchport port-security
 switchport port-security maximum 1
 switchport port-security mac-address sticky
 switchport port-security violation shutdown
 spanning-tree portfast
 spanning-tree bpduguard &lt;span class="nb"&gt;enable&lt;/span&gt;
 storm-control broadcast level 1.00
 ip dhcp snooping limit rate 15
errdisable recovery cause psecure-violation
errdisable recovery cause bpduguard
errdisable recovery interval 300
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;DHCP snooping and DAI go on at the VLAN level, with trust set on the uplinks only. Forget the uplink trust and the VLAN loses its DHCP server, so do this in a maintenance window.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;ip dhcp snooping
ip dhcp snooping vlan 100-102
ip arp inspection vlan 100-102
interface Port-channel1
 ip dhcp snooping trust
 ip arp inspection trust
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;DAI validates ARP against the snooping binding table. A camera with a static address never asked a DHCP server, so it has no binding, and its ARP is dropped as soon as DAI turns on. Half the cameras go dark. The cleaner fix is a DHCP reservation for every camera and door controller. For devices that must stay static, add an ARP ACL and attach it to the VLAN.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;arp access-list STATIC-DEVICES
 permit ip host 192.0.2.50 mac host 0011.2233.4455
ip arp inspection filter STATIC-DEVICES vlan 100
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;A static source binding for the device does the same job. Whichever you pick, build the list from the camera inventory before enabling DAI, and watch &lt;code&gt;show ip arp inspection statistics&lt;/code&gt; for drops afterward.&lt;/p&gt;

&lt;h2&gt;
  
  
  Unused ports
&lt;/h2&gt;

&lt;p&gt;Every port without a labelled purpose goes into a VLAN that has no SVI and no routing, gets shut down, and carries a description that says it is unused.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;interface range GigabitEthernet1/0/24, GigabitEthernet1/0/30 - 48
 switchport mode access
 switchport access vlan 666
 description UNUSED
 shutdown
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Backups and firmware
&lt;/h2&gt;

&lt;p&gt;Save the config after each change set with &lt;code&gt;copy running-config startup-config&lt;/code&gt;, then pull a copy off the switch on a schedule.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;copy running-config scp://backup@192.0.2.40/sw01.cfg
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Keep the backup server out of the camera VLANs, and diff each copy against the last one. For firmware, check the target release against Cisco’s security advisories before you deploy, and use install mode so a failed upgrade can roll back.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;install add file flash:&lt;span class="o"&gt;&amp;lt;&lt;/span&gt;image&lt;span class="o"&gt;&amp;gt;&lt;/span&gt;.bin activate commit
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Checking your work
&lt;/h2&gt;

&lt;p&gt;The &lt;a href="https://hans.study/tools/switch-config-audit/" rel="noopener noreferrer"&gt;switch config audit tool&lt;/a&gt; reads a pasted running config and flags the gaps against this baseline, and the &lt;a href="https://hans.study/switch-config-audit/" rel="noopener noreferrer"&gt;switch config audit&lt;/a&gt; is the remote review across a whole estate. Hands-on practice is in the &lt;a href="https://hans.study/training/network-and-windows-hardening/" rel="noopener noreferrer"&gt;Network and Windows Hardening&lt;/a&gt; course.&lt;/p&gt;

&lt;p&gt;Book a working session at &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;/book/&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>networking</category>
      <category>cisco</category>
      <category>security</category>
      <category>cybersecurity</category>
    </item>
    <item>
      <title>Aruba CX AOS-CX Hardening: A Practical Baseline</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Mon, 09 Nov 2026 23:30:57 +0000</pubDate>
      <link>https://dev.to/hansstudy/aruba-cx-aos-cx-hardening-a-practical-baseline-258g</link>
      <guid>https://dev.to/hansstudy/aruba-cx-aos-cx-hardening-a-practical-baseline-258g</guid>
      <description>&lt;p&gt;A hardened AOS-CX baseline puts SSH on the management VRF behind an allow-list, authenticates logins centrally with a local fallback, moves SNMP to v3 only, and authenticates NTP. On the access ports it adds port security, BPDU guard or loop protection, DHCP snooping, and ARP inspection, with unused ports shut down in a blackhole VLAN. As on any platform, static-IP cameras need explicit handling before ARP inspection goes on. Plan for that first.&lt;/p&gt;

&lt;p&gt;The syntax below was checked against the HPE AOS-CX Command-line interface Reference Guide for 8100 and 8360 Switch Series, published February 2026. Other series such as the 6200 and 6300 share most of it, but confirm against the guide for your model and release. Addresses are from the 192.0.2.0/24 documentation range, and every secret is a placeholder. The Cisco equivalent is in &lt;a href="https://hans.study/hardening-cisco-catalyst-switches-ios-xe-baseline/" rel="noopener noreferrer"&gt;hardening Cisco Catalyst switches&lt;/a&gt;, and the audit side is in the &lt;a href="https://hans.study/standards-guidance/switch-configuration-audit-checklist-for-security-networks/" rel="noopener noreferrer"&gt;switch configuration audit checklist&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Management plane
&lt;/h2&gt;

&lt;p&gt;SSH is disabled until the admin password is set, and it is enabled per VRF. It will not start without it. Put it on the management VRF, restrict who can connect, and turn off Telnet and the web interface where you do not use them. The CLI reference states that the Telnet server is disabled by default, so the explicit line below is a guard against someone enabling it later.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;ssh server vrf mgmt
ssh server allow-list
 ip 192.0.2.0/24
 &lt;span class="nb"&gt;enable&lt;/span&gt;
ssh maximum-auth-attempts 3
no telnet server vrf mgmt
no https-server vrf mgmt
session-timeout 10
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Enabling the allow-list restarts the SSH servers and ends active sessions, and an enabled list with no entries blocks everyone. Apply it from the console or from a session inside the permitted range. Test it. Keep the HTTPS server only if you use the REST API or web interface, and then keep it on the management VRF.&lt;/p&gt;

&lt;p&gt;Authentication comes next. Create named users in the built-in groups, point SSH and console logins at TACACS+ with local last, and test the fallback before you leave the room. AOS-CX stores local passwords hashed, so there is no separate secret type to choose.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;user ops1 group administrators password
tacacs-server host 192.0.2.20 key plaintext &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt; vrf mgmt
aaa authentication login ssh group tacacs &lt;span class="kd"&gt;local&lt;/span&gt;
aaa authentication login console group tacacs &lt;span class="kd"&gt;local&lt;/span&gt;
aaa authentication limit-login-attempts 5 lockout-time 300
banner motd ^
This system is private property of &lt;span class="o"&gt;[&lt;/span&gt;ORGANIZATION&lt;span class="o"&gt;]&lt;/span&gt;. Authorized use only.
Activity is logged and monitored.
^
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The guide warns that command authorization has no fail-through, so a wrong TACACS+ authorization setup can leave a switch unable to run commands. Always end the group list with local, and prove the setup on a lab switch first.&lt;/p&gt;

&lt;h2&gt;
  
  
  Time, SNMP, and logging
&lt;/h2&gt;

&lt;p&gt;NTP can be tied to a trusted key so the switch only syncs with a source that proves itself, which protects log timestamps from a spoofed time server on the camera VLAN. SNMPv3 users get authentication and privacy, and the switch can be told to drop v1 and v2c entirely. Syslog goes out over the management VRF.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;ntp authentication-key 10 sha1 &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt; trusted
ntp authentication
ntp server 192.0.2.30 key 10
ntp vrf mgmt
ntp &lt;span class="nb"&gt;enable&lt;/span&gt;
snmp-server snmpv3-only
snmpv3 security-level auth-privacy
snmpv3 user monitor auth sha256 auth-pass plaintext &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt; priv aes priv-pass plaintext &lt;span class="o"&gt;&amp;lt;&lt;/span&gt;strong-secret&lt;span class="o"&gt;&amp;gt;&lt;/span&gt; access-level ro
logging 192.0.2.10 severity info vrf mgmt
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The guide notes that a # character in an NTP key stops synchronization once authentication is on, so choose keys without one. The logging command also supports TCP and TLS transports, and TLS is the better choice when your collector can present a certificate.&lt;/p&gt;

&lt;h2&gt;
  
  
  Control plane basics
&lt;/h2&gt;

&lt;p&gt;AOS-CX applies a predefined control plane policing policy named default at first startup. Review it with &lt;code&gt;show copp-policy&lt;/code&gt; rather than rewriting it. Leave it alone. Rate limits on ingress ICMP and broadcast traffic on edge ports add a second layer, and the management plane is already limited by the VRF and allow-list above.&lt;/p&gt;

&lt;h2&gt;
  
  
  Access ports
&lt;/h2&gt;

&lt;p&gt;Edge ports get an access VLAN, edge port type, BPDU guard, loop protection, a broadcast rate limit, and port security, all applied the same way to every port that faces a camera, reader, or workstation. The client limit defaults to 1. Sticky learning keeps the first MAC the port sees. A violation shuts the port down, and auto-recovery brings it back after a timer.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;spanning-tree bpdu-guard timeout 300
port-access port-security &lt;span class="nb"&gt;enable&lt;/span&gt;
interface 1/1/1
 no routing
 vlan access 100
 spanning-tree port-type admin-edge
 spanning-tree bpdu-guard
 loop-protect
 rate-limit broadcast 1 percent
 port-access port-security &lt;span class="nb"&gt;enable&lt;/span&gt;
 port-access security violation action shutdown
 port-access security violation action shutdown auto-recovery &lt;span class="nb"&gt;enable&lt;/span&gt;
 port-access security violation action shutdown recovery-timer 300
 port-access port-security
  client-limit 1
  sticky-learn &lt;span class="nb"&gt;enable&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Port security with one client catches an unmanaged switch added for extra drops, while BPDU guard catches a managed switch or a loop through one. Loop protection is a separate mechanism that sends its own control packets, so it also catches loops that never involve spanning tree.&lt;/p&gt;

&lt;p&gt;DHCP snooping is enabled globally and per VLAN, and only the ports that face the real DHCP server, usually the uplinks or the LAG toward the distribution layer, are trusted. Everything else stays untrusted. From release 10.14 the command is &lt;code&gt;dhcp-snooping&lt;/code&gt;, and earlier releases use &lt;code&gt;dhcpv4-snooping&lt;/code&gt;. ARP inspection is enabled per VLAN on the 8100 and 8360 from 10.12.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;dhcp-snooping
vlan 100
 dhcp-snooping
 arp inspection
interface lag 1
 dhcp-snooping trust
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;ARP inspection discards ARP packets that carry an invalid IP-to-MAC binding, so a camera with a static address and no lease is at risk of being dropped. That is the trap. Convert cameras and door controllers to DHCP reservations where you can. For devices that must stay static, the guide documents &lt;code&gt;ipv4 source-binding&lt;/code&gt;, which adds a static entry to the binding database.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;ipv4 source-binding 100 192.0.2.50 00:11:22:33:44:55 1/1/5
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The guide documents how to enable ARP inspection and view its trust state per interface, but I could not find a command to set a port as trusted in it. Check the security guide for your release, and keep uplinks in mind when you test, because a wrongly untrusted uplink drops ARP between switches.&lt;/p&gt;

&lt;h2&gt;
  
  
  Unused ports
&lt;/h2&gt;

&lt;p&gt;Unused ports go into a VLAN with no routed interface, get a description, and are shut down.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;vlan 666
 name BLACKHOLE
interface 1/1/24
 no routing
 vlan access 666
 description UNUSED
 shutdown
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Backups and firmware
&lt;/h2&gt;

&lt;p&gt;Copy the running configuration to a remote server over SFTP or SCP on a schedule, using the management VRF, and diff each copy against the last. The guide lists the copy forms for running configuration and checkpoints.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight cisco_ios"&gt;&lt;code&gt;copy running-config sftp://backup@192.0.2.40/sw01.cfg vrf mgmt
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;For firmware, read the release notes and the HPE security bulletins for the target version before upgrading. AOS-CX keeps primary and secondary images, so stage the new image on the secondary partition and keep the known-good one until the switch has run clean for a day.&lt;/p&gt;

&lt;h2&gt;
  
  
  Checking your work
&lt;/h2&gt;

&lt;p&gt;The &lt;a href="https://hans.study/tools/switch-config-audit/" rel="noopener noreferrer"&gt;switch config audit tool&lt;/a&gt; checks a configuration you paste or upload against a published checklist. The &lt;a href="https://hans.study/switch-config-audit/" rel="noopener noreferrer"&gt;switch config audit&lt;/a&gt; is the remote review across a whole estate, and the &lt;a href="https://hans.study/learning/network-hardening-level-1/" rel="noopener noreferrer"&gt;Network Hardening Level 1&lt;/a&gt; course and the &lt;a href="https://hans.study/training/network-and-windows-hardening/" rel="noopener noreferrer"&gt;Network and Windows Hardening&lt;/a&gt; training cover the reasoning behind each control.&lt;/p&gt;

&lt;p&gt;Book a working session at &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;/book/&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>networking</category>
      <category>security</category>
      <category>cybersecurity</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>What counts as Specified Information under CPCSC?</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Sun, 08 Nov 2026 18:58:19 +0000</pubDate>
      <link>https://dev.to/hansstudy/what-counts-as-specified-information-under-cpcsc-1ba0</link>
      <guid>https://dev.to/hansstudy/what-counts-as-specified-information-under-cpcsc-1ba0</guid>
      <description>&lt;p&gt;Specified Information is government of Canada information that must be protected when a non-government organization handles, processes, or stores it. That is PSPC’s definition in its Level 1 scoping guide. Which of your documents qualify is set by your contract, not by a general list.&lt;/p&gt;

&lt;h2&gt;
  
  
  The reasoning
&lt;/h2&gt;

&lt;p&gt;The official pages don’t give a catalogue of document types. PSPC tells suppliers to use their contract, security clauses, and any handling instructions to confirm what counts, and then to identify every asset and environment that has it. So the work starts in the paperwork, not on the file server.&lt;/p&gt;

&lt;p&gt;On the hub I describe it as sensitive but unclassified government information that a contract identifies as needing safeguarding on supplier systems. The usual candidates are drawings, statements of work, schedules, and technical data. That list is my experience, not PSPC’s wording, and the contract decides each case.&lt;/p&gt;

&lt;p&gt;One drawing can be Specified Information on one job and ordinary correspondence on another. Two purchase orders for the same part can come out differently.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where it ends up
&lt;/h2&gt;

&lt;p&gt;PSPC’s scoping guide says in-scope assets are the ones that process, store, or transmit it. Its examples include computers, mobile devices, file shares, cloud storage, email, servers, backup media, and printers or scanners that can store information. The guide also describes information that is received, created, edited, printed, emailed, transferred, backed up, and destroyed.&lt;/p&gt;

&lt;p&gt;That list matters because the copy you know about is rarely the only one. Echoes in backups, sync clients, and downloads folders count.&lt;/p&gt;

&lt;p&gt;The same logic covers what you create. A routing sheet, a first-article inspection report, or a quote that restates details from a customer’s drawing can carry the same handling duty, depending on the clause. PSPC’s guide describes information that is created, edited, and generated by an automated system, not only information that arrives from outside.&lt;/p&gt;

&lt;h2&gt;
  
  
  Edge cases
&lt;/h2&gt;

&lt;p&gt;If a clause is vague, ask the contracting authority or your prime what is meant. Get the answer in writing.&lt;/p&gt;

&lt;p&gt;If you can’t tell, treat the material as Specified Information until you can. Wrongly protecting a file costs little, and wrongly ignoring one costs a lot.&lt;/p&gt;

&lt;p&gt;Controlled Goods data is a separate regime. A drawing may be both, and meeting one doesn’t satisfy the other.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do this week
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; Collect the security clauses and any handling instructions from every defence-related contract.&lt;/li&gt;
&lt;li&gt; Write down, per contract, what the paperwork says is Specified Information.&lt;/li&gt;
&lt;li&gt; Trace each item through 5 stages, from arrival to rest, work, leaving, and echoes.&lt;/li&gt;
&lt;li&gt; Mark every system it touches, which becomes your boundary.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The &lt;a href="https://hans.study/cpcsc/scoping/" rel="noopener noreferrer"&gt;scoping page&lt;/a&gt; walks through the trace and the ways to shrink the footprint, and &lt;a href="https://hans.study/cpcsc/does-it-apply/" rel="noopener noreferrer"&gt;does it apply&lt;/a&gt; covers the yes-or-no test. &lt;a href="https://hans.study/cpcsc-bytes/" rel="noopener noreferrer"&gt;CPCSC Bytes&lt;/a&gt; has shorter explainers.&lt;/p&gt;

&lt;p&gt;If you would like help reading your clauses and mapping the data, &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;book a scoping call&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>cybersecurity</category>
      <category>compliance</category>
      <category>canada</category>
      <category>security</category>
    </item>
    <item>
      <title>How long CPCSC Level 2 readiness takes a small shop</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Sat, 07 Nov 2026 18:59:26 +0000</pubDate>
      <link>https://dev.to/hansstudy/how-long-cpcsc-level-2-readiness-takes-a-small-shop-10dm</link>
      <guid>https://dev.to/hansstudy/how-long-cpcsc-level-2-readiness-takes-a-small-shop-10dm</guid>
      <description>&lt;p&gt;By my estimate, a small shop with a contained network and current equipment needs about 6 months of part-time effort to get ready for Level 2. A flat network with shared admin accounts and no logging needs closer to 18 months, plus some capital. Those are my numbers from the work involved, not figures PSPC has published.&lt;/p&gt;

&lt;h2&gt;
  
  
  The reasoning
&lt;/h2&gt;

&lt;p&gt;Level 2 assesses all 98 ITSP.10.171 requirements, through a certification body accredited by the Standards Council of Canada, with an annual affirmation between assessments. That is a long way from Level 1’s 13 requirements and a yearly self-attestation.&lt;/p&gt;

&lt;p&gt;The time goes into governance more than technology. A firewall can be bought and installed in a week. Written policies that people follow, a system security plan that matches the running network, access reviews, log reviews, and one piece of evidence for every requirement take months to build and months more to prove through habit.&lt;/p&gt;

&lt;p&gt;Scope sets the calendar. If Specified Information is confined to an enclave, you prepare the enclave. If it has spread across every laptop and mailbox, the same 98 requirements apply to the whole company. The &lt;a href="https://hans.study/cpcsc/scoping/" rel="noopener noreferrer"&gt;scoping page&lt;/a&gt; covers the shrinking steps.&lt;/p&gt;

&lt;h2&gt;
  
  
  What PSPC has published
&lt;/h2&gt;

&lt;p&gt;PSPC’s September 29, 2026 program overview says Levels 2 and 3 are under development and gives no dates. Its April 2026 announcement planned Level 2 for select contracts from spring 2027, and its earlier timeline page read April 2027 to March 2028. As of my last check in early October, the assessment methodology, the accredited certification body list, and the rules for gaps found at assessment were not published.&lt;/p&gt;

&lt;p&gt;PSPC also says suppliers will first need to show they meet the new Level 1 controls. Level 1 comes first, then, and it is the opening step of any Level 2 plan I would draw.&lt;/p&gt;

&lt;p&gt;A rough split of my 6-month case looks like this. The first 6 to 8 weeks go to scoping and the Level 1 foundation. The next stretch goes to policies and the system security plan, written to match what the network really does. The last stretch goes to collecting evidence, running your own rehearsal, and fixing what it finds. A flat network adds a rebuild phase in front of all that, and the rebuild is where the capital goes.&lt;/p&gt;

&lt;h2&gt;
  
  
  Edge cases
&lt;/h2&gt;

&lt;p&gt;Open items at assessment are the biggest unknown. CMMC allows limited plans of action with a 180-day closeout, but no published Canadian position exists, so don’t assume the same here.&lt;/p&gt;

&lt;p&gt;If your prime never asks for Level 2, you may not need it for years. The contract names the level, so read the clauses before you commit a budget.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do this week
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; Confirm which contracts carry Specified Information clauses.&lt;/li&gt;
&lt;li&gt; Map where that data lives and draw a boundary.&lt;/li&gt;
&lt;li&gt; Run a gap check against the &lt;a href="https://hans.study/cpcsc/requirements/" rel="noopener noreferrer"&gt;98 requirements&lt;/a&gt; and note which are policy gaps and which are technical.&lt;/li&gt;
&lt;li&gt; Complete &lt;a href="https://hans.study/cpcsc/level-1/" rel="noopener noreferrer"&gt;Level 1&lt;/a&gt;, since it is everyone’s first step.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The &lt;a href="https://hans.study/cpcsc/level-2/" rel="noopener noreferrer"&gt;Level 2 page&lt;/a&gt; tracks PSPC’s timing changes as they happen, and &lt;a href="https://hans.study/cpcsc-bytes/" rel="noopener noreferrer"&gt;CPCSC Bytes&lt;/a&gt; carries shorter notes.&lt;/p&gt;

&lt;p&gt;If you want to size your own timeline, &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;book a scoping call&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>cybersecurity</category>
      <category>compliance</category>
      <category>canada</category>
      <category>security</category>
    </item>
    <item>
      <title>Is MFA required for CPCSC Level 1?</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Fri, 06 Nov 2026 20:30:42 +0000</pubDate>
      <link>https://dev.to/hansstudy/is-mfa-required-for-cpcsc-level-1-35k4</link>
      <guid>https://dev.to/hansstudy/is-mfa-required-for-cpcsc-level-1-35k4</guid>
      <description>&lt;p&gt;Yes. Multifactor authentication is requirement 03.05.03, one of the 13 CPCSC Level 1 requirements. PSPC’s how-to page says it is required for privileged accounts and for systems that store Specified Information, and I would plan for every account that touches those systems.&lt;/p&gt;

&lt;h2&gt;
  
  
  The reasoning
&lt;/h2&gt;

&lt;p&gt;PSPC’s overview lists MFA as control 7 of the Level 1 set, under identification and authentication. Its how-to page words the duty as MFA for privileged accounts and for systems that store Specified Information. ITSP.10.171 and PSPC’s assessment criteria cover privileged and non-privileged accounts, which is the wider reading.&lt;/p&gt;

&lt;p&gt;I plan for the wider reading. The criteria are what you sign against when you attest, and a gap on ordinary user accounts that can open the drawings is hard to defend. MFA is cheap to turn on and expensive to explain away.&lt;/p&gt;

&lt;p&gt;Shops coming from CMMC are often surprised by this. The American Level 1 has no MFA requirement, so passing it on passwords alone says nothing about CPCSC Level 1.&lt;/p&gt;

&lt;h2&gt;
  
  
  Edge cases
&lt;/h2&gt;

&lt;p&gt;Which systems are covered depends on your boundary. If Specified Information lives in an enclave file share plus email, MFA covers those logins and the admin accounts that manage them. Draw the boundary first, using the &lt;a href="https://hans.study/cpcsc/scoping/" rel="noopener noreferrer"&gt;scoping page&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Shared logins are a problem. MFA tied to a shared shop-floor account doesn’t identify anyone, and requirement 03.05.01 asks for unique identification.&lt;/p&gt;

&lt;p&gt;On method, authenticator apps work and hardware keys are the option I push. SMS is the weakest choice, and I would avoid it for admin accounts.&lt;/p&gt;

&lt;p&gt;Evidence is simple. PSPC’s list of examples includes screenshots of MFA configuration, so a dated screenshot per system is enough to start.&lt;/p&gt;

&lt;p&gt;Remote access deserves its own look. A VPN or remote desktop gateway that reaches the enclave is a privileged path in practice, even when the account behind it is an ordinary one. If MFA is on the Microsoft 365 sign-in but not on the VPN, the front door is locked and the side door isn’t. Fix both before you attest.&lt;/p&gt;

&lt;p&gt;The cost of getting this wrong is mostly an attestation problem. Level 1 has no routine evidence review, and PSPC reserves the right to look, so the signature is what carries the weight. I would rather a shop delay its attestation by two weeks to turn MFA on than sign and hope.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do this week
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; List every system that stores or can reach Specified Information, including email and remote access.&lt;/li&gt;
&lt;li&gt; List every privileged account on them, and then every ordinary user account.&lt;/li&gt;
&lt;li&gt; Turn on MFA for the admin accounts first, then everyone else.&lt;/li&gt;
&lt;li&gt; Capture a dated screenshot of each configuration and file it under 03.05.03.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The &lt;a href="https://hans.study/cpcsc/level-1/" rel="noopener noreferrer"&gt;Level 1 page&lt;/a&gt; has the other 12 requirements, and &lt;a href="https://hans.study/cpcsc-bytes/" rel="noopener noreferrer"&gt;CPCSC Bytes&lt;/a&gt; has short notes on the rest.&lt;/p&gt;

&lt;p&gt;To check MFA coverage against your own boundary, &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;book a scoping call&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>cybersecurity</category>
      <category>compliance</category>
      <category>canada</category>
      <category>security</category>
    </item>
    <item>
      <title>CPCSC Level 1 evidence a prime will ask for</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Thu, 05 Nov 2026 21:10:36 +0000</pubDate>
      <link>https://dev.to/hansstudy/cpcsc-level-1-evidence-a-prime-will-ask-for-3la3</link>
      <guid>https://dev.to/hansstudy/cpcsc-level-1-evidence-a-prime-will-ask-for-3la3</guid>
      <description>&lt;p&gt;Expect two kinds of request. The first is proof that you attested, which is the results page and the expiry date confirmed in your CanadaBuys supplier profile. The second is whatever supports the 13 answers, which at Level 1 you keep yourself and show on request.&lt;/p&gt;

&lt;h2&gt;
  
  
  The reasoning
&lt;/h2&gt;

&lt;p&gt;Level 1 is an annual self-assessment against 13 of the 98 ITSP.10.171 requirements. PSPC’s Level 1 page tells suppliers to keep evidence for the duration of the attestation cycle, or at least one year, and lists examples such as account and device lists, access review notes, security policies, training records, patch and sanitization logs, visitor logs, firewall settings, and screenshots of MFA configuration.&lt;/p&gt;

&lt;p&gt;PSPC also says you must provide proof of self-attestation and its expiry date in CanadaBuys, and again with a bid on a contract that requires Level 1. That proof is the part nobody can skip. The rest sits in your files until someone asks.&lt;/p&gt;

&lt;h2&gt;
  
  
  What a prime is likely to ask
&lt;/h2&gt;

&lt;p&gt;PSPC doesn’t set a prime’s questionnaire, so this part is my expectation. Primes already send supplier assurance questions about insurance and quality, and cyber status is a natural addition. I would expect these, roughly in this order:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;  Your results page with its expiry date&lt;/li&gt;
&lt;li&gt;  Which contracts, and which systems, your boundary covers&lt;/li&gt;
&lt;li&gt;  A named person who signed the attestation&lt;/li&gt;
&lt;li&gt;  Evidence for the requirements a prime cares about most, usually MFA, patching, and who has access&lt;/li&gt;
&lt;li&gt;  A statement of what happens when your environment changes&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;None of that is a published prime checklist. Read your contract’s clauses and ask the prime what form it wants.&lt;/p&gt;

&lt;h2&gt;
  
  
  Edge cases
&lt;/h2&gt;

&lt;p&gt;The attestation is a signed representation in a federal procurement setting, so an answer you can’t back up is a risk to you rather than a formality. If a requirement isn’t met yet, fix it before you attest. Level 1 gaps are usually cheap to close.&lt;/p&gt;

&lt;p&gt;Your boundary matters here too. Evidence only has to cover the systems where Specified Information lives, so a tight scope means fewer artifacts. The &lt;a href="https://hans.study/cpcsc/scoping/" rel="noopener noreferrer"&gt;scoping page&lt;/a&gt; covers how to draw it.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do this week
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt; Make a folder per requirement identifier, 13 in all.&lt;/li&gt;
&lt;li&gt; Put one dated artifact in each, such as a screenshot, an export, or a signed policy.&lt;/li&gt;
&lt;li&gt; Save the PSPC results page next to a note of the expiry date, and calendar the renewal a month early.&lt;/li&gt;
&lt;li&gt; Draw the boundary on one page and date it.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The &lt;a href="https://hans.study/cpcsc/level-1/" rel="noopener noreferrer"&gt;Level 1 page&lt;/a&gt; lists the 13 requirements in plain language, and the &lt;a href="https://hans.study/cpcsc/templates/level-1-checklist/" rel="noopener noreferrer"&gt;Level 1 checklist&lt;/a&gt; pairs each with the evidence an assessor’s methods would reach for. More short explainers live in &lt;a href="https://hans.study/cpcsc-bytes/" rel="noopener noreferrer"&gt;CPCSC Bytes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;To go through your own evidence folder with someone who reads these requirements daily, &lt;a href="https://hans.study/book/" rel="noopener noreferrer"&gt;book a scoping call&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>cybersecurity</category>
      <category>compliance</category>
      <category>canada</category>
      <category>security</category>
    </item>
    <item>
      <title>Your MSP's remote tool is inside your compliance scope</title>
      <dc:creator>Hans Study</dc:creator>
      <pubDate>Wed, 04 Nov 2026 20:55:51 +0000</pubDate>
      <link>https://dev.to/hansstudy/your-msps-remote-tool-is-inside-your-compliance-scope-36m5</link>
      <guid>https://dev.to/hansstudy/your-msps-remote-tool-is-inside-your-compliance-scope-36m5</guid>
      <description>&lt;p&gt;The MSP is the one party in a CPCSC program that doesn't sign anything and still ends up in the assessment. It isn't a defense supplier. Its client is. But if it administers the systems where the client's Specified Information lives, its access, its tooling, and its personnel are inside the client's scope, and the assessor will look at them.&lt;/p&gt;

&lt;p&gt;Most MSP relationships weren't built with that in mind.&lt;/p&gt;

&lt;h2&gt;
  
  
  The bench with a remote tool on every machine
&lt;/h2&gt;

&lt;p&gt;I've had MSPs tell me "this is how we do it" while their whole bench had a remote tool on every machine.&lt;/p&gt;

&lt;p&gt;That is very hard to defend against the access control and maintenance requirements, because it isn't role-based anything. Does every technician need access to everything? Almost never. And do the people who hold access have the screening the contract requires?&lt;/p&gt;

&lt;p&gt;There is also a security argument that has nothing to do with compliance. I've seen remote support tools compromised and used to wipe fleets. The tool that lets the MSP fix your server is the same tool an attacker wants.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I'd build instead
&lt;/h2&gt;

&lt;p&gt;Put remote access through a jump host with MFA on the session. 03.07.05, non-local maintenance, requires MFA for it. The technician authenticates to the jump host, the jump host is the only path into the enclave, and the jump host is where the logging happens.&lt;/p&gt;

&lt;p&gt;Keep administrative work on separate admin workstations, and never work under an admin account day to day. 03.14.09 sets a stricter bar for those workstations.&lt;/p&gt;

&lt;p&gt;If a remote support tool stays, turn on session recording from the moment of login.&lt;/p&gt;

&lt;p&gt;The assessor's questions are ordinary. When was the server last touched? By whom? Show the login and logout. Show the patch window and what was applied. Screenshots and a basic central log server answer all of it. Nobody is asking for a SIEM.&lt;/p&gt;

&lt;h2&gt;
  
  
  Who assesses the MSP's work
&lt;/h2&gt;

&lt;p&gt;Your IT provider shouldn't be assessing its own work. Some won't tell on themselves.&lt;/p&gt;

&lt;p&gt;So keep the checking at arm's length. Run your own vulnerability scans, with proper tools. Audit switch and firewall configurations with something like rConfig or Nipper. If the MSP says "we do vulnerability management for you," or "our tool says you're secure," verify it with your own scan and your own read of the configs.&lt;/p&gt;

&lt;p&gt;Two red flags. A provider that hides behind "proprietary" or refuses configuration access. And an MSP running a sales-oriented assessment tool and calling it a security assessment. It may tick a box, and it is still a sales tool.&lt;/p&gt;

&lt;h2&gt;
  
  
  Put it in a matrix
&lt;/h2&gt;

&lt;p&gt;"They handle our IT" is not a control description. The program forces a written &lt;a href="https://hans.study/cpcsc/templates/shared-responsibility-matrix/" rel="noopener noreferrer"&gt;responsibility matrix&lt;/a&gt;: requirement by requirement, who operates the control, who holds the evidence, and what deliverable arrives on what cadence. Sample rows from the one I use:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Patch management: MSP operates, MSP holds evidence, patch report monthly&lt;/li&gt;
&lt;li&gt;Firewall rule set: MSP operates, rule review quarterly, client keeps a copy&lt;/li&gt;
&lt;li&gt;Log collection and review: MSP operates, weekly digest to a client mailbox&lt;/li&gt;
&lt;li&gt;Backup and restore test: MSP operates, restore test log quarterly&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If the MSP runs the control, the MSP produces the evidence, on a schedule, as a contract deliverable. Have that conversation early and in writing. It surfaces 2 uncomfortable discoveries at a survivable moment. Some MSPs can't support this work, and 2026 is a better time to learn that than mid-assessment. And the MSP's own security posture is part of your risk story, because its remote access into your enclave is one of your attack paths.&lt;/p&gt;

</description>
      <category>cybersecurity</category>
      <category>compliance</category>
      <category>sysadmin</category>
      <category>canada</category>
    </item>
  </channel>
</rss>
