SOME BORING NETWORK
ENGINEERING INTERVIEW
QUESTIONS
AND HOW TO REPLACE THEM WITH SMARTER CHOICES
KAM AGAHIAN
DIRECTOR, NETWORK INFRASTRUCTURE- UNIVERSITY OF SOUTHERN CALIFORNIA (USC)
CCIE#25341 X2
NE INTERVIEWS
• We will cover
• Fundamentals of network engineering interview topics
• The overlooked criteria
• Examples of what you don’t want to ask/be-asked
• Smarter choices and how they help?
THE PILLARS OF BEING A TRADITIONAL NE
• TCP/UDP/IP
• IGP
• BGP
• Other areas of expertise: MPLS – Data Center – Automation
• Voice, Security, Systems (e.g. Linux) – a little bit untraditional
(THE) GREAT EXPECTATIONS
• TCP/UDP/IP
• TCP, UDP and IP headers and how they’re used?
• TCP session establishment and ending process
• TCP congestion control mechanism and slow start process
• TCP/UDP advantages and disadvantages
• Perhaps some higher layer troubleshooting tools such as trace route
(THE) GREAT EXPECTATIONS
• IGP of choice
• OSPF
• OSPF theories and messaging model
• Single and multi area designs and challenges
• LSA types (perhaps v4 vs v6), interface types, timers
• Filtering, summarization and optimization
• ISIS
• ISIS theories and messaging model
• Single and multi area designs and challenges
• LSP types, reasons behind TLV types and metric challenges
• Single and multi topology and IPv6 interactions
• Filtering, summarization and optimization
(THE) GREAT EXPECTATIONS
• BGP – Sometimes depends on the line of business
• BGP
• Why BGP?
• BGP theories including the state machine and messaging
• BGP attributes and how they help in various scenarios such as traffic manipulation
• Summarization, filtering and optimization (e.g. “newer” methods such as PIC, 4B ASN)
• Global BGP and its challenges
• BGP in enterprises and its challenges
• BGP in data center designs and its challenges
• BGP in carriers and its challenges
NE INTERVIEWS…
• Were traditionally designed to measure
candidates’ technical depth
• Barely covers only one aspect of being a successful NE
• Inception of the boring questions…
WHAT WE TEND TO OVERLOOK
• Room for proper leveling
• To have flexibility to hire at a lower or higher levels
• To make sure the candidate is at the level you are hiring
• To be able to compare two good candidates
WHAT WE TEND TO OVERLOOK
• Room for candidates’ creativity
• To give room to candidates to imagine, design and build
• To gauge candidates’ willingness and passion to even engage
in discussions
• To measure candidates’ thought process beyond cold hard
technical solutions
SAMPLE QUESTIONS
• Ready to explore some of the topics?
TCP/UDP/IP
• What is the difference between TCP and
UDP?
TCP/UDP/IP
• What are the differences between TCP and
UDP?
• Intention: Mostly whether the candidate is
familiar with reliability at layer 4.
TCP/UDP/IP
• What are the differences between TCP and
UDP?
• Major issues:
• Interviewers tend to break off the chase after hearing the
word “reliable” or “ACK” or “Handshake”
• Can be superficially memorized without deep understanding
• No room for creativity or leveling (i.e. it’s flat)
TCP/UDP/IP – THE PLANE SCENARIO
• Let’s design a fictious aviation tracking
system
• We want our fleet of planes to report their
locations and other vital parameters once every
5 minutes. Let’s say we had a choice to use TCP
or UDP which one would you pick and why?
TCP/UDP/IP – THE PLANE SCENARIO
• The plane scenario
• The candidate needs to:
• To understand stakes, risks and definition of success
• To evaluate options: Reliability vs Simplicity
• To articulate reasoning
• To defend ideas (bonus point: soft skills)
• Perhaps expose various hidden layers in their design (e.g. cost, other
options etc.)
IGP
• What are the differences between OSPF and
ISIS?
IGP
• What are the differences between OSPF and
ISIS? Or Which one do you like better?!
• Intention: Whether the candidate has any
experience with both protocols – maybe?
IGP
• What are the differences between OSPF and
ISIS?
• Major issues:
• The two protocols very few similarities!
• Messaging, leveling, timers, applications, traffic engineering, design considerations and
many more are all different. But they’re both Link State!
• Can be superficially memorized without deep understanding
• No room for creativity or leveling (i.e. it’s flat)
IGP
• Hidden risk:
• Biased interviewing
• The candidate can name 10 differences without hitting THE
difference that the interviewer has in mind. They are all correct
but still not the correct answer.
• Then follows a series of hints which later during the debrief
will be perceived and described as “hand holding”.
IGP – THE EXPERIMENTAL ROUTING SCENARIO
• Let’s be creative for a bit and design a
fictious routing protocol
• We have limited time and resources available
and need to create a link state protocol to
support an experimental type of IP. Would start
you work based off OSPF or ISIS and why?
IGP - THE EXPERIMENTAL ROUTING SCENARIO
• The Experimental Routing Protocol scenario
• The candidate needs to:
• To understand the limitations
• To identify the key to success: Fewest number of changes
• To articulate reasons
• Perhaps expose various hidden layers in their design (the IPv6
experience, phases, other options etc.)
• Maybe they can challenge the interviewer's imitations by going after BGP
and optimizing it.
BGP
• Case 1: Explain the BGP decision process
• Intention: Whether the candidate is aware
of the steps in the right order.
BGP
• Case 1: Explain the BGP decision process
• Major issues:
• More of a memory test
• No room for creativity or leveling (i.e. it’s flat)
BGP
• Case 2: Which one is well known, optional,
mandatory, discretionary etc.?
• Major issues:
• More of a memory test
• No room for creativity or leveling (i.e. it’s flat)
BGP – THE EGB DEVELOPMENT CASE
• With limited time and resources you are
developing an open source experimental
version of BGP.
What BGP attributes would you develop in
the first phase? And why?
BGP - THE EGB DEVELOPMENT CASE
• Experimental EGP
• The candidate needs to:
• Understand the concept of well-known and mandatory attributes
• Know what attributes are well-known and mandatory
• Understand BGP beyond simple lab setups
• Understand the risks and be able to articulate an example
BGP
• Case 3: How do you influence inbound (or
outbound) traffic:
• Intention: Whether the candidate can apply
the BGP attributes to a semi-real world
scenario.
BGP
• Case 3: How do you influence inbound (or
outbound) traffic.
• Major issues:
• Very classic answers that can be memorized
in minutes
• No room for creativity or leveling (i.e. it’s flat)
BGP– THE MED/AS-PATH SCENARIO
• In order to influence inbound Internet traffic into
the network a network engineer is proposing use
of the MED.
What is the point she is missing?
BGP
• The MED/AS-PATH scenario
• The candidate needs to:
• To understand limitations and definition of success
• Know how the MED works; a lot of details to understand
• Know why the transitive and non-transitive attributes exist
• Know why the well-known mandatory attributes exist
• Understand BGP beyond simple lab setups.
• Understand the risks and be able to articulate an example
BGP
• The MED/AS-PATH scenario
• Also leaves room for further layers and thinking outside the box:
• What if due to lack of expertise we prefer not to mess with BGP
attributes?
• Advertise more specific prefixes
• Do not advertise
Q/A
