Sunday, March 14, 2010

Too Many Directions, Too Much to Think About

July 31, 2009 by Greg Ferro · 2 Comments 

Too many dir­ec­tions at once

It’s been a busy time. I have been set­tling into a new con­sult­ing role as a Network Architect. Starting is always tough, there are a lot of new people and new polit­ics to under­stand. Plus the bur­den of under­stand­ing the pro­ced­ures because every busi­ness does the same thing dif­fer­ently (and no, I don’t under­stand how that is possible).

This role is almost purely a design and archi­tec­ture role. That means to receive the out­put from the Business Analyst (a work of fic­tion and guess­work that includes estim­ates on how much the pro­ject will cost) and turn it into a Detailed Design. The Detailed Design will then go to an engin­eer for implementation.

So far, nor­mal for most people who do this kind of work.

Diversity

Over the years, I have become com­pet­ent at writ­ing doc­u­ments, find­ing and com­mu­nic­at­ing with the vari­ous people, devel­op­ing a con­sensus for a design that is accept­able to everyone.

942B7948-C7FA-482D-B614-87F1F39490D9.jpg

The thing that I am find­ing most chal­len­ging is the diversity of tech­no­lo­gies. This week has seen myself work­ing on pro­jects that have too many vendors and too many tech­no­lo­gies and my head can’t hold all the data. The fol­low­ing is an over­view of the tech­no­lo­gies /​ products I have been involved with this week:

  • Blue Coat ProxySG, ProxyAV doing HTTP proxy, scan­ning and con­tent filtering.
  • Cisco ACE load bal­an­cer deployment
  • Packeteer traffic shaping
  • BGP Site-​​of-​​Origin /​ EIGRP MPLS Site-​​of-​​Origin
  • Nokia/​Checkpoint fire­walls to build a DMZ
  • Cisco ASA /​ PIX fire­walls for the same DMZ.
  • PVLANs for another DMZ, using trunks on mul­tiple switches with VMWare serv­ers. Meaning con­fig­ur­a­tion of Cisco switches, VMWare and the firewalls.
  • cre­at­ing a new site includ­ing IP phones for a PABX over a IPSec VPN over Radio network.
  • New VLANs/​SVI versus sec­ond­ary addresses

you get the idea, I am sure.

Its Detailed, alright

If I was work­ing on any one of these, I think I would be OK. But doing all of these in single week I am los­ing my abil­ity to com­plete designs. The point of pre­par­ing a Design Document is to cap­ture the detailed inform­a­tion like IP Addressing, VLAN num­bers as well as the big pic­ture that sets the busi­ness require­ments, and keep an eye on what prob­lem is so that pro­ject actu­ally deliv­ers a result.

A design has two parts — the cre­at­ive writ­ing part that cov­ers the busi­ness and pro­ject deliv­ery and tech­nical writ­ing part con­tains highly detailed and pre­cise inform­a­tion that the implment­a­tion team will need. I find that doing both of these is hard, espe­cially the cre­at­ive part.

I can be hap­pily work­ing on one design, only to take ques­tions on another —- instant brain wreck and con­cen­tra­tion loss.

Is there an CCIE angle to this ?

Well yes. One of the things that the CCIE exam taught me is how to func­tion with stuffed full head. Most men­dic­ants tell you about how their head is over­loaded and not sure how they will cope on exam day. For what it’s worth, I get that feel­ing most days. It doesn’t get any easier.

But the exam does teach you about your­self, and more or less, makes you find a way to cope with large amounts of diverse information.

And that is part of the reason that many suc­cess­ful CCIE’s go on to senior roles, and are at fore­front of new devel­op­ments in most com­pan­ies. Sure, tech­nical know­ledge and com­pet­ence is import­ant, but it’s not the only thing. The abil­ity to “grasp and men­tally hold” the entire net­work for a large com­pany, includ­ing the higher lay­ers such a Servers, Applications, Databases requires practice.

A good engin­eer knows tech­no­logy. But the men­tal dis­cip­line and prac­tice to gather, hold, sort that data, and then extract mean­ing from it and being able to com­pre­hend a wide range of tech­no­logy means you have a chance to be a great engineer.

I guess I’m still work­ing on it. Sigh.

Please rate this post:

  Why Rate Posts?
1 Star - It\\\'s Crud2 Stars - It\\\'s Tosh3 Stars - Something\\\'s missing4 Stars - Needs works5 Stars - Good Enough6 Stars - Good7 Stars - Excellent8 Stars - Brilliant9 Stars - Astonishing10 Stars - Awesomely Godlike? (7 votes, average: 8.71 out of 10)
Loading ... Loading ...

Comments

2 Responses to “Too Many Directions, Too Much to Think About”
  1. Ethan says:

    A good engin­eer knows tech­no­logy. But the men­tal dis­cip­line and prac­tice to gather, hold, sort that data, and then extract mean­ing from it and being able to com­pre­hend a wide range of tech­no­logy means you have a chance to be a great engineer.

    Well said. Bravo.

    • I’m won­der­ing what your imple­ment­a­tion engin­eers think about your work. I’ve real­ised that it’s hard to make a dis­tinct line between design and imple­ment­a­tion phase. May be some require­ments were missed or imple­ment­a­tion didn’t go as planned and you have to imple­ment a work­around which may affect your design.

      And I saw a lot of vari­ous inter­rac­tions between design and ser­vice deliv­ery people:

      Smart Designer with a lot of field exp(CCIE) + Dumb Implementators = Design include detailed con­fig­ur­a­tions of the hard­ware with detailed phys­ical lay­outs. Field engin­eers simply con­nect devices as described, put IOS and con­figs. In case of prob­lem, design engin­eer log in remotely and invest­ig­ate what’s going on.

      Smart Designer + Smart Implementators = Designer starts giv­ing more gen­eral con­figs, ser­vice deliv­ery engin­eer can make IP alloc­a­tion by him­self, etc… Troubleshooting is done together. Usually smart field engin­eers go out from field work espe­cially people with fam­il­ies as they don’t have time for kids if they travel too much.

      Dumb Designer + Smart Implementators = Designs will be very high-​​level. Like cloud to cloud con­nec­ted with one link without any details. Service Delivery will eat a lot of shit deal­ing with the imple­ment­a­tion. Basically they have to make a new design (we can call it low-​​level, but it’s com­pletely dif­fer­ent kind of job). It’s very easy to talk to cus­tom­ers and draw clouds in Visio. But if there are no good imple­ment­a­tion engin­eers, dumb designer will be screwed (or pro­moted to high man­age­ment as it usu­ally happens)

      Dumb Designer + Dumb Implementator = Fail! :)

Speak Your Mind

Tell us what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!