Influence

A Product Manager's biggest strength and challenge is the ability to influence. As a PM,  you need to be able to get work done from folks that do not report to you like engineers, developers, sales and marketing. You need to possess the ability to hold people accountable without real authority over them. You have to indirectly control them through influence.

So how do you influence :
1. Work towards the same goal : meeting customer needs. All employees are working to satisfy customer needs, remind them of this ultimate business goal.
2. Improve the process that shows how time and cost efficient we can be as team
3. Build a personal relationship, make the team like and understand you. Congratulate them on a big release, applaud their efforts in front of the team.
4. Provide an impact scenario; if we don't deliver on time, we are losing a major customer. Make them realize that they are working for a greater cause. Most engineers never get insight into the big picture. By exposing them to a real world problem solved by their solution motivates and makes them understand the gravity of the situation. Instill reality, make them feel accountable and responsible.

It's all about the power to influence, to motivate and to inspire. And this influences, motivates and inspires me to wake up and go to work everyday and build a world class product with an amazing team.

Networking protocols

Today we will talk about protocols that allow networking and I will keep updating this as I think of relevant protocols.

Starting with layer 2 
Spanning tree protocol(STP) is a protocol that allows for path redundancy while eliminating undesired loops in the network and this is accomplished by blocking certain links.
How does this work?
1. A root bridge is elected ( switch with lowest MAC address if Bridge ID same for all)
2. Root ports on a non-root bridge ( shortest route to root bridge)
3. Elect designated port on each device and that will forward traffic, rest of the link will be blocked

Transparent Interconnect of lots of links (TRILL) is a  replacement for STP
Why? STP wastes bandwidth by blocking certain links.
Trill brings routing intelligence to layer 2, link state protocol runs on RBridges(devices that support TRILL, usually switches). The link state protocol provides the topology of the RBridges and the surrounding links, so each RBridge computes shorted path to the RBridge for data transmission.

There are 3 types of routing protocols

• Interior Gateway Protocol using link state  routing protocol (OSPF, IS-IS)
• Interior Gateway Protocol using Distance vector routing protocol ( RIPv2, IGRP and EIGRP)
• Exterior Gateway Protocol ( BGP)

Open Shortest Path First (OSPF)
This is a link state protocol that gathers topology information from the neighboring nodes and determines the shortest path to destination using Dijkstra's algorithm. OSPF is a preferred for more than 4-5 routers. This protocol will send out the routing table only when it detects a topology change unlike RIP that sends the table every 30 seconds. This protocol is used mainly in campus enterprise solutions. The way OSPF works is, it selects a Designated Router(DR) and a Backup DR and based on talking to neighbors through a hello and acknowledgment with calculate shortest path with determine route. OSPFv3 has been designed for IPv6 traffic. ISIS is independent of IPv4 or IPv6 traffic.

Intermediate State to Intermediate State (IS-IS)
ISIS is a link state protocol used mainly in service provide environments. ISIS supports more routers than OSPF so preferred more by Internet Service Provider(ISP). ISIS also appends a TLV(type length value) to its packets which is useful in ISP networks.

Routing Information Protocol (RIP)
This protocol is based on the distance vector routing protocol. Routes are based on hop count. RIP can support a max. of 16 hops. The way this protocol works is, the RIP router will send a REQUEST broadcast message to all the RIP enabled routers in its network. Then those routers reply with a RESPONSE routing table that the RIP router will use to update routing information. RIP is much older and there is a delay to convergence. Convergence is a when all the RIP enabled routers have the same topology information as the network they belong to. This mechanism is preferable when the routers know the topology well and its a small network. RIP sends routing table updates every 30 secondss which wastes a lot of bandwidth.

Border Gateway Protocol(BGP)
BGP is an exterior gateway protocol that used path vector protocol. This is a layer 4 protocol, sits on top of TCP. BGP in a simple way is a routing protocol that makes the internet work. All the large internet service providers use BGP, its the most important internet protocol.

Multi-protocol Label Switching (MPLS)
Traditional switch will look at the next hop and transport packet to that. The MPLS mechanism looks the destination address and selects a pre-determines path to the destination by putting a label on the packet. Like a courier service, postman does not open the package to see where it needs to be delivered just looks at the label for the destination address. This mechanism is protocol independent( DSL or IP ) and provides QoS. ( VPN does not provide this)

VLANs

Lets talk about VLANs today, to understand VLAN you must know what a LAN is.
LAN is a local area network - it is a group of devices connected with a cable. Every packet goes to each device in a LAN. As the devices grow, collision can occur since a wire might already have a packet sent by another device. You can physically separate the device with routers but then you cant share resources also its physically connected and you will need to plug and unplug each time you want to connect certain devices each other. This is where VLAN comes in.

VLAN is a virtual LAN, a group of devices interact with each other like they are in the same domain. VLAN is a subset of the LAN. So a switch with 48 ports can be segmented such as ports 1- 24 are in one VLAN and ports 25-48 are in another VLAN.

Benefits of VLAN:
1. Scalability - Single switch to multiple segments - 2 floors you can have 1 switch with 2 vlans
2. Better performance than segmenting with routers
3. Partitioning for special projects - if only certain people need access to a confidential file they will be on one VLAN
4. Flexibility - its easy to move devices from one VLAN to another through configurations

Types of VLANs:
1. Static or port based VLAN - there is a VLAN ID tagged on the packet
2. Protocol based VLAN
3. Subnet based VLAN

OSI Model

Lets start with the basics, what is the OSI model and what does it do. OSI stands for Open Systems Interconnect and this model talks about how a network communicates and transmits data. OSI model constitutes of 7 layers - Physical, Data Link , Network, Transport, Session, Presentation and Application. Best way to remember this  - (P)lease (D)o (N)ot (T)hrow (S)tuffed (P)izza (A)way

Physical Layer ( layer 1 )- This provides the hardware for communication between devices. Ex - copper cable, optical cable, wireless 802.11N technology, patch panel.

Data Link Layer( layer 2 )- This is your link layer, provides a link between the connected devices. Ex - switch

Network Layer( layer 3 ) - This layer deals with the logical addressing ( IPv4, IPv6), determines best route for data transmission( static or dynamic routing) and fragmentation ( breaking the message into smaller pieces so it can fit into the frame)

Transport Layer( layer 4 )- This deals with reliable data transmission( TCP,UDP), flow control, sequencing( making sure all the packets are in the correct order) and segmentation ( segments the messages into smaller peices so as to suit the requirements of the network layer)

Session Layer ( layer 5 )- This layers opens up communication between devices to start data transmission

Presentation Layer ( layer 6 ) - In this layer ASCII data is converted into readable format at the destination port and vice verson on the source port

Application Layer ( layer 7) - This layer interacts with the applications. Ex - outlook, http, ftp

 Diagram to help understand : 

In short :)