Skip to main content

thinking about server upgrades

current setup is working but thinking about upgrades. also might be procrastinating dealing with trading losses.

current setup
#

home server:

  • dell poweredge r730
  • dual xeon e5-2680 v4 (28 cores total)
  • 128gb ram
  • 2x 1tb nvme ssd
  • runs all my backtesting, monitoring, databases

network:

  • ubiquiti dream machine pro
  • 10gbe to server
  • 1gbe internet (spectrum business)

cost: paid $2,800 for server on ebay, $400 for networking

works fine. no latency issues for my current strategies (1-5 minute bars, not HFT).

what i’m considering
#

colocation in chicago

  • rent rack space near CME datacenter
  • put dedicated server there
  • sub-millisecond latency to exchange

cost: $150-300/month + $2k for server

question: do i actually need this? my strategies aren’t latency-sensitive. not doing market making or arbitrage.

probably not worth it until i’m profitable enough to afford burning $3k/year on colo.

storage upgrade
#

backtesting generates TB of data. currently at 800gb used on 1tb drives.

options:

  • add more nvme drives ($400 for 2x2tb)
  • migrate old backtest data to spinning rust ($150 for 8tb HDD)

leaning toward HDD for old data. don’t need nvme speeds for 2020 backtests i’ll never run again.

the real question
#

am i thinking about server upgrades because i need them? or because it’s easier than thinking about being down $60k?

probably the second one.

servers are concrete problems with concrete solutions. trading losses are abstract pain with unclear path forward.

what i actually need
#

less server upgrades, more strategy improvements.

my hardware can handle 100x my current trading volume. the bottleneck isn’t CPU or network latency.

the bottleneck is my strategies suck.

but also
#

the dell r730 is sick AF and i love having a server rack in my closet. LEDs everywhere. sounds like a jet engine. probably using $100/month in electricity.

worth it? questionable. cool factor? 100%.


3:22am. gonna price out the HDD storage upgrade but probably won’t buy anything. need to focus on trading not hardware.

-AK

Related

data feed comparison - polygon vs alpha vantage
your backtest is only as good as your data. spent $200/month on polygon.io and holy shit the difference vs free data. what i was using (bad) # alpha vantage free tier
grafana monitoring setup for algo trading
can’t fix what you don’t measure. built a grafana dashboard to track everything my algos are doing. why i need this # when strategies shit the bed, i need to know immediately. not 2 hours later when i check my phone and see -$5k.