State of the Lab 2016

I last shared my Homelab in 2014 and after updates over the last two years my lab barely resembles the one from two years ago. These State of the Lab updates will become a more frequent series where I share not only the hardware of the lab, but also various software components that I’m using or testing in the lab. The primary focus of any lab is for experimentation and learning so this should be an interesting series. Especially considering I have over 35 VMs running in the lab for everything from Plex to Exchange 2016 to Team Foundation Server 2015.

Current Lab

The lab currently consists of the following hardware:

APC NetShelter CX Noise Isolating Server Cabinet

  • Size: 24U

HP Storage Works 1/8 G2 Tape Library

  • LTO-4 (800GB per tape)
  • 8 Tape Capacity
  • SAS

Lenovo RD630 (decommissioned)

  • CPU: 1 x Intel Xeon E5-2620 2Ghz
  • RAM: 48GB
  • RAID: LSI 9260-8i w/ 512MB Cache & BBU
  • HDD: 4 x 146GB 15K RPM SAS
  • IMPI: ThinkServer Management Premium Module

Dell R620

  • OS: Windows Server 2012 R2 Datacenter w/ Hyper-V
  • CPU: 2 x Intel Xeon E5-2690 2.9Ghz w/ HT (32 Logical Cores)
  • RAM: 64GB
  • RAID 1 (Internal): PERC H710P
  • RAID 2 (External): LSI 9280-8e (Connected to SA120)
  • HDD: 4 x 900GB 15K SAS (RAID 5 + 1 Hot Spare)
  • HBA: LSI 9200-8e (Connected to HP Storage Works Tape Library)

Lenovo SA120

  • HDD: 4 x 1TB 7200RPM (RAID 5)
  • HDD: 2 x 5TB 7200RPM (RAID 0)

HP DL360 Gen8

  • OS: VMware ESXi 6.0U2
  • CPU: 2 x Intel Xeon E5-2670 2.6Ghz w/ HT (32 Logical Cores)
  • RAM: 256GB
  • RAID: HP P420i
  • HDD: 7 x 900 GB 15K SAS (RAID 5)
  • SSD: 1 x 1.2TB (vFRC)

APC SMT2200RM2U Rackmount UPS

  • Addon: APC9631 Network Management Card
  • Runtime: ~1 hour battery runtime @ ~500W load

Misc Hardware

  • Switch: Cisco SG500-28P (24 Port POE L3 Managed Switch)
  • Wireless: Unifi AP-AC Pro
  • Firewall: Cisco Meraki MX 64

Next post I’ll share some insight onto what I’m running on this hardware and various things on my list to throw up in the lab.

What room is there for Man in automated reality?

Lately the news is flooded with a few consistent overarching issues when it comes to the labor market. An often cited issue is the higher than desired unemployment numbers, high labor participation dropout rates, and the push to increase the minimum wage across the US. The recent recession played a heavy role in the loss of jobs, but what is keeping them from coming back?

The Worker

Ever since the dawn of the Industrial Revolution there has been a question of how well humans can compete against machines. The mechanization of the early days often made people fear that they would no longer be able to keep up. On the surface this fear is well placed. After all, something that previously may have taken a group of workers weeks or months to complete was now doable in a few days by a single worker. Although the fear was well placed this doomsday view never came to fruition. There never was a tempering in the demand for human labor, even as human population grew exponentially. The biggest shift became an overarching demand for skilled human labor. The machine could do a lot to empower the worker, but it couldn’t replace him.

I have felt for years that the modern revolution is vastly different from the dawn of the Industrial Revolution. Not only are the machines finally capable of replacing a large share of workers, but that the workers will find it impossible to keep pace. The machine no longer needs to be driven by an operator at all times and increasingly can drive itself. This creates a fundamental problem for capitalistic societies where workers are another raw material like any other.

The Challenge

The big challenge as automation comes to dominate the workforce is how do we adjust our definition of work and where people fit. In an increasingly automated world we no longer have a need for an increasing share of the population in an economically productive manner. As the cost of labor increases, via methods like increased minimum wages and other means, the push to automate will continue to increase. How do we approach a future where the worker not only has to compete with labor in various places around the world, but also with automation.

Fix Education

I feel the most important push against the machines is to ensure that future labor force participants are equipped to offer value where machines currently can’t (information transformation aka decisions). Our current method of education focuses heavily on memorization and repetition as it has for the last few hundred years. Humans have already lost the battle of calculation speed and information retrieval to machines. There is little value in being able to recite the encyclopedia when it can be queried by the average person in seconds.

Students will need to be trained how to transform information into something of value. How to derive meaning from the information presented to them while leveraging the power of computing to work on ever larger data sets in ever more creative ways. We need to end the teaching of rote memorization and teach critical thinking on a deep and expansive level to give future workers a fair chance.

Reevaluate Taxes

Automation is so attractive because of the structure of our economic and tax systems. Our current system is structured with the understanding that lower taxes on businesses and the wealthy leads to increased job creation and a larger pool of employed persons. Without automation some of this may prove to be factual (although I’m not convinced it is factual). With extremely low taxes business is given a strong incentive to automate and to employ as few people as possible. Business can afford the enormous cost of automation while reaping the benefits of lower expenses (both in saved labor costs and a lower tax burden).

We need to reevaluate the basic structure of our tax system. A solidifying of safety nets for workers would be a basic start. Corporations should be taxed at a higher rate and workers should be given certain protections. It should be very cheap, or free, for workers to retrain to keep their skills in demand. The burden of education should also be lowered. A worker education insurance program similar to unemployment insurance should exist to assist workers with education and living cost to enable transition.

Conclusion

We need to reassess our understanding of work in a future where people are becoming increasingly obsolete. Certain lines of work will be primarily human for the foreseeable future, such as Barbers and Lawyers, but many lines are increasingly being replaced by better and better machines. Automation eats at the workforce from the bottom up putting most workers at risk eventually.

The most straightforward solution is to address education to refocus it toward the demands of the present. We can’t afford to focus education of our youth on the expectation that they will spend their lives working in the same profession. We have to ensure that workers can adjust from one field that is no longer needed to new fields that are in demand with the lowest burden possible. This should be paid for by the business community as a trade-off for increased automation. Ultimately, we have to ensure that business can’t externalize the cost of human labor back onto society.

Home Lab

Every technology professional needs a solid technology lab. Without this in place you’ll fall behind in relevant technology changes because you won’t be able to properly experiment. Being able to spin up a machine running the latest version of Windows Server or CentOS can be invaluable.

My Current Lab

The core of the lab is composed of a Dell R620 ESXi host running a variety of VMs for almost everything. It has enough power to enable me to add as many VMs as I need with the biggest limitation being in raw disk space. Additionally, I use a Cisco Smart Switch to power the networking between the variety of the hardwired devices. For Wireless LAN I use a UniFi Wireless Access Point to extend the network to laptops, phones, tablets, and guests.

Home Lab Server Rack
Home Lab Server Rack

Switch: Cisco SG200-26 Smart Switch
Wireless Access Point: UniFi AP Pro (2.4Ghz & 5Ghz Simultaneous Dual Band)
Primary Server (bottom): Dell R620
CPU: Dual Intel Xeon E5-2690 @ 2.9Ghz w/ 8 cores + HyperThreading (32 logical cores)
RAM: 64GB DDR3 ECC RAM
HDD: 2 x 600GB, 1x 900GB 2.5″ SAS 10K RPM (RAID 5)
Dell Perc H710 Hardware RAID controller w/ BBU
ESXi 5.5 booted via internal USB header

Secondary Server (top): Lenovo RD630
CPU: Single Intel Xeon E5-2620 @ 2.0 Ghz w/ 6 cores + HyperThreading (12 logical cores)
RAM: 4GB DDR3 ECC RAM
HDD: 3 x 500GB 2.5″ WD Enterprise SATA (RAID 5)
HDD: 1 x 160GB 2.5″ SATA Drive (for ISOs and VM templates)
ThinkRaid 700 Hardware RAID controller w/ BBU
ESXi 5.1 booted via internal USB header

This stack should be powerful enough to last into the foreseeable future. The R620 has over 15 active VMs running various servers for everything from SIP Phone service (3CX) to powering a local DropBox like service (using OwnCloud) and firewall (pfSense). With power usage under 130W I’m able to get fantastic performance per watt out of this machine.

Future blog posts will offer insights on how to properly setup a variety of local services in your lab that I find very useful.

Credit Scoring Basics

I decided to write this blog article on credit scoring after noticing the number of questions I got about this topic from family and friends. Demystifying credit scoring is an important part of your financial health and this article will get you started. I will be compiling a whole series on credit that will cover various topics in-depth.

What Is Credit Scoring?

Simply put, credit scoring is one method banks use to measure credit risk in consumer lending. Since credit scoring is all about risk prediction it is important to remember that the score is only focused on ensuring banks can adequately price risk. It is important to keep this in mind as at times actions that may increase your credit score may not be in your best interest.

A Word of Caution

The most popular model for Credit Scoring was invented by Fair Issac and is marketed under the FICO brand. Since FICO scoring is the most popular model used in the consumer lending space the advice in this article will be focused on information geared towards the FICO scoring models. Alternative models, often regarded to as FAKO, won’t be discussed due to there being no real way of knowing what the model consists of.

Although the majority of consumer lending institutions use FICO scoring when determining your interest rate they are allowed to customize the model. As a result, your credit score can vary from institution to institution even without any material difference in your credit report.

What Is A Credit Score Composed Of?

According to Fair Issac a FICO Score is composed of the following weighted components:

Weight Category Note
35% Payment History Late payments and collections
30% Amounts Owed Your utilization – How close your balances are to your credit limits.
15% Account Age Your Average Age of Accounts.
10% New Credit How often you’ve sought credit.
10% Types of Credit How diverse your types of credit are – such as having a mortgage and credit cards.

FICO scores range from 350 – 850, with a score above 700 generally considered as “good” credit.

Payment History

This is the most important component of a credit score. It is also the hardest to change as the only cure for a poor payment history is time. As such, the most important thing to do to obtain and maintain good credit is to make on-time payments to any lenders you have. This seems to go without saying, but many people don’t understand the disastrous impact a single late credit card payment can have on their credit score and borrowing costs.

According to the Fair Credit Reporting Act negative information needs to be removed in the time frames specified in the below chart:

Negative Information Type Time to Removal Note
Late Account (Non-Medical) 7 Years The time is based on “date of first delinquency” not when it appeared on the report.
Medical Collections 10 Years
Public Records 10 Years Chapter 13 Bankruptcy is 7 years.
Tax Liens 15 Years

Positive information for active accounts can remain on your report indefinitely and closed paid accounts can remain for up to 10 years.

Utilization (Amounts Owed)

The second most important factor affecting your FICO score is your credit utilization. FICO scoring models look at both your overall utilization as well as your per card utilization. Ideally, you want to keep you utilization lower than 10% overall and individually to maximize your score.

There are a few important factors to know about utilization. The first is that it is not historically tracked so there is no need to follow this number until you are actively seeking credit. If you have high utilization and you’re actively seeking credit the easiest and biggest score boost will come from lowering your utilization. Another important factor for utilization is to be aware that it is tracked by the statement balance reported on your credit card statement. This is very important because if you’re like me and pay off your card balance every month you won’t get any extra points for doing so. FICO doesn’t care if you pay off 100% of your balance monthly or only pay the card minimum. Although this seems counter intuitive it makes sense. Remember, FICO is all about predicting your risk of default in the future. In that regard, having a large amount of liabilities are a risk. A job loss could theoretically ruin you.

Average Age of Accounts

Average age of accounts is a measure of how old your credit scoring data is. This is not the same age as your credit file itself. The older your credit scoring data the bigger the boost you’ll get from this component. Of course, the whole reason for this is to give more weight to reports that span a larger share of the user’s credit life.

Inquiries (New Credit)

Every time you actively seek new credit (such as applying for a credit card or mortgage) you’ll take a slight credit score hit. This hit will come from increased inquiries and can drop your score as much as 10%. All inquiries sit on your credit report for 2 years.

Due to the impact of applying for new accounts it is best to hold off on any new credit applications for 6-12 months before seeking any major loans (such as a mortgage). Failure to do so may result in a lower score at the time of application forcing you to incur higher borrowing costs or be declined altogether.

Account Mix (Types of Credit)

FICO scoring prefers you to have a mix of credit types, such as mortgage loans and credit cards. This is one of the smallest factors in your credit score and I don’t recommend seeking the perfect mix. You shouldn’t borrow money, especially in large fixed installment loans (like a mortgage), solely for credit scoring benefits. Without the right mix you may never reach a 850 score, but if you manage the other components you’ll have a score high enough to keep your borrowing costs low.

Resources

Resource Note
My FICO Fair Issac MyFICO consumer site
FICO Community MyFICO Community Forums
Credit Boards Credit Boards credit information forum