About Zack Systems

Zack Logo I was chief technical officer (CTO) of Zack Systems from February of 2000 until we shut the company down in March 2002. Zack built a software + hardware appliance platform designed to put high-value user-facing services into any IP network. With the Zack platform any network provider could offer virus scanning, spam filtering, language translation, popup blocking, and other in-network services on an ala carte or package basis. All this was achieved without requiring any client-side or server-side software install.

While at Zack I:

  • Led the design and development of Zack's platform
  • Shipped 3 products
  • Directly hired 11 engineers, plus 15 indirectly
  • Helped author and file 7 broad patent applications
  • Represented the company's technologies and engineering skill to customers and potential investors.
  • Helped raise $8.5MM in venture capital
  • Helped close $500K in purchase orders

I designed Zack's scalable, reliable, secure platform and led its implementation. Seven broad patent applications were filed based upon the innovations we developed. At Zack's largest, we had 30 engineers working on the product and over 50 employees. We achieved successful full-time test deployments including 100% of traffic flow in two ISPs and smaller trial deployments in half a dozen others.

Ultimately, Zack ran out of money because our customers (carriers and network service providers) had run out of money. Right as our product reached the market the market collapsed and investment capital dried up. Nevertheless, I'm convinced that we will eventually see ala carte edge services like virus scanning and spam filtering become the essential ingredient for profitability in many ISP/NSP/carrier contexts, just as we have seen network-based services like call waiting and caller id become the engine driving local telco profits.

(See the links below for archived Zack PDF marketing and technical deliverables.)

The Technology

The Zack platform was capable of examining and modifying traffic flowing across a network in real time, including web pages, email messages, and other protocols. Applications written to the platform could add, subtract, monitor, and modify network traffic. Typical added latency was 10-20 milliseconds -- far below the level of human perception. Zack appliances were transparent to end users except for any services offered. Application programming interfaces (APIs) simplified the creation of new services and products and streamlined integration with billing and provisioning systems. An integrated, consistent carrier-accessible (and potentially user-accessible) interface was provided to control any number of applications running on the platform.

Zack appliances were clusterable for near-linear scalability and high reliability. When installed in conjunction with a layer 4 switch the Zack platform supported an "n+1" reliability model, where by sizing a cluster one machine larger than the expected traffic load the loss of any one machine in the cluster could be handled with no degradation in service. Extensive provisions were incorporated into the Zack platform for fault detection, automatic failure isolation, and graceful degradation when faced with hardware failures or software bugs.

The appliance hardware itself was a high quality commodity 1U rack-mount x86 server unit with modest RAM and disk requirements. A hardened Linux kernel was used to keep software cost down. The entire platform could be upgraded or downgraded in the field via a cryptographically secure packaging system.

See also:   Zack Systems Executive Summary
Zack Edge Application Server Brochure
Zack Platform Technology Whitepaper


Please note that most of this site was last updated in 2003.

 - Home page
 - Overview
 - Consulting
 - Zack Systems
 - Photography
 - Mobiles
Code / Docs
 - Palm Bus Schedule
 - Small perl scripts
 - Old stuff
About Jason
 - Contact Info
 - Tools
 - Drop Box
 - Weblog
  copyright © 2003 Jason D. Campbell, All Rights Reserved.
  File /work/zack/index.mason, updated 13 May 2003