Skip to content
Log in Sign up My account

ENSLD Cram Sheet

The facts and trade-offs most worth re-reading the night before 300-420 ENSLD, one section per blueprint domain. ENSLD is a design exam, so most lines here are "when to choose X" rather than commands.

1.0 Advanced Addressing and Routing Solutions

Addressing (1.1)

  • Allocate on bit boundaries so each site, region and function summarises to one prefix. Leave room to grow; do not pack blocks tightly.
  • Use separate blocks for loopbacks, point-to-point links, user VLANs, voice, infrastructure and management.
  • IPv6: a /48 per site is the common plan, a /64 per VLAN (always, for SLAAC), a /127 or /64 on point-to-point links. A /48 holds 65,536 /64 subnets.
  • Align IPv6 boundaries on nibbles (multiples of 4 bits) so prefixes are easy to read and summarise.
  • ULA is fc00::/7 (in practice fd00::/8), not routed on the internet. 2001:db8::/32 is for documentation.

IGP design traits (1.2, 1.3, 1.4)

OSPF IS-IS EIGRP
Algorithm Link state Link state Advanced distance vector (DUAL)
Hierarchy Areas; area 0 backbone; ABRs split areas Levels: L1 (intra-area), L2 (backbone), L1/L2 routers join them No built-in areas; hierarchy comes from summarisation and stubs
Area border On a router (ABR) On the link: area is per router, adjacencies form per level Summary at any router
Transport IP protocol 89 Directly over Layer 2 (CLNS), not IP IP protocol 88
Scaling limit is LSDB size, SPF frequency, ABR count Same, with fewer LSA types and a flat backbone Query scope and stuck-in-active risk
Limit the blast radius with Multiple areas, summarisation, stub/NSSA areas, prefix-suppression L1 areas with default via attached bit, route summarisation or leaking at L1/L2 Summarise, eigrp stub, distribute-list, tune queries
Fast convergence BFD, SPF/LSA throttling, i-SPF, LFA BFD, SPF throttling, LFA Feasible successor (no recompute), BFD
Metric Cost (reference bandwidth) Narrow (max 63 per link) or wide metric Bandwidth and delay
Best fit Mixed vendor enterprise, well understood Large flat backbones, SD-Access underlay, service providers Cisco-only, hub-and-spoke, simple summarisation
  • OSPF: all areas must connect to area 0 (virtual link is a temporary repair, not a design). Summarise on the ABR (inter-area) and ASBR (external). Keep areas to a manageable router count and keep the ABR count per area low.
  • OSPF stub area removes external LSAs. Totally stubby removes inter-area LSAs as well. NSSA allows an ASBR inside a stub-like area (Type 7 converted to Type 5 at the ABR).
  • IS-IS: an L1 router only knows its area and sends everything else to the nearest L1/L2 router (the attached bit). Backbone must be a contiguous set of L2-capable routers.
  • EIGRP: a summary at the distribution layer stops queries and hides flaps. Stub spokes are never asked for queries. EIGRP named mode and wide metrics help with high-speed links.
  • Passive-interface on user-facing links and authentication on IGP adjacencies are the baseline for secure routing.

BGP (1.5)

Attribute Scope Influences Prefer
Weight Local to the router, Cisco only Outbound Higher
Local preference Whole AS (iBGP) Outbound Higher (default 100)
AS-path length Between ASes Inbound (prepending) Shorter
Origin Transitive Tie-break IGP over EGP over incomplete
MED Sent to the neighbouring AS only Inbound (suggestion) Lower
eBGP over iBGP Per path Tie-break eBGP
IGP metric to next hop Per path Tie-break Lower
Community Tag Policy signalling, for example provider local-pref n/a
  • Best path order to remember: weight, local preference, locally originated, AS-path, origin, MED, eBGP over iBGP, lowest IGP metric to next hop, then tie-breakers.
  • Outbound control: local preference. Inbound control: prepending, MED (only with one neighbouring AS), more-specific prefixes and provider communities.
  • Route reflectors remove the full-mesh requirement. Use at least two, in different locations, with a shared cluster ID. Loop prevention uses ORIGINATOR_ID and CLUSTER_LIST. Confederations are the alternative.
  • Do not become a transit AS: outbound filter should permit only your own prefixes (prefix-list and an empty or own-origin AS-path filter).
  • Load sharing: maximum-paths for eBGP multipath, maximum-paths ibgp for iBGP. Paths must tie on weight, local preference, AS-path length, origin, MED and IGP metric. Different upstream ASes need bestpath as-path multipath-relax.
  • Address families: MP-BGP carries IPv4, IPv6, VPNv4 and others over one session. A neighbour must be activated per family. IPv6 can run over an IPv4 session and the reverse.
  • Filtering tools: prefix-list (match prefix and length), AS-path access-list (regex on path), route-map (set and match), community lists, maximum-prefix as a safety limit.

IPv6 migration (1.6)

Strategy What it is Choose when Cost
Dual stack (1.6.b) IPv4 and IPv6 run natively on the same devices Default for campus and DC; devices support both Double the operational state (routing, ACLs, monitoring)
Overlay tunnel (1.6.a) IPv6 inside IPv4 (manual GRE, IPv6IP, 6to4, ISATAP, 6rd) IPv6 islands across an IPv4-only core MTU reduction, no native performance, tunnel endpoints to manage
Translation (1.6.c) NAT64 with DNS64, NAT46, 464XLAT IPv6-only clients must reach IPv4-only servers State on the translator, breaks end-to-end address use
  • NAT64 uses a /96 prefix; the well-known prefix is 64:ff9b::/96. DNS64 synthesises AAAA records for IPv4-only names.
  • Tunnels are an interim step. Native dual stack is the target.

Free account required for the rest

The rest of this page is for members. You're seeing a free sample. A free account unlocks everything: every question bank, the timed exams, the cram sheets, the simulations and the lab files.

Create a free account Log in