DigiTech Magazine - US Summer 2017 - Page 8


By Ben Grinnell
When Netflix moved to the cloud , it released a “ chaos monkey ” to wreak havoc on its system . The chaos monkey is a script meant to mimic the behavior of “ a wild monkey with a weapon in your data center ( or cloud region ) to randomly shoot down instances and chew through cables .” 1 The idea was to cause intentional failures in a controlled environment where Netflix ’ s experts could comfortably monitor and repair its infrastructure .
During lulls in operations , the Netflix team either increased the activity of the chaos monkey or decreased the tolerances on the many metrics it recorded on its apps and infrastructure . This proactive approach to identifying problems was not new . Toyota ’ s plant managers have been known to tighten tolerances when employee reports of assembly line problems drop in order to detect ever smaller weaknesses . At Netflix , these self-inflicted disasters enabled the company to identify weak links and build up resiliency so it could continue serving customers when real failures occurred .
Modern IT solutions must be able to continue functioning when their own wild animals break into the system . By working closely with other developers and service owners , developers build a deep understanding of how their applications and infrastructure perform through a culture of measurement , continuous improvement and learning . This is the only way to survive the inevitable disasters .
Innovation and constant improvement are essential to the long-term success of enterprise IT . Over the last couple of decades , however , many leaders decided that IT was not a core capability they wanted to worry about . They outsourced it , optimizing for cost and reliability . That decision built up an entire industry and led many cost-focused organizations to delay modernizing their requirements . IT fell behind , so when those initial contracts with business process outsourcing ( BPO ) providers were relet , the IT Refresh program was born .
In this program , the new tenders invited the industry to bid to refresh and then run their IT . Most prioritized cost . They saw the price of IT hardware fall , and they assumed that IT expenses overall would drop with it . Tenders frequently expected their suppliers to maintain rates at existing levels and assumed that the supplier would recoup IT Refresh costs over the life of the new contract .
In the short term , this model appealed to many CIOs because it was an easy sell throughout the organization . Over the first 18 months of the contract , everyone got new technology and modernized services , and the cost of IT held steady . Unfortunately , this model didn ’ t account for the increasing complexity and advancement of IT , users ’ growing expectations of basic services , and the rising IT Refresh costs that accompanied it . This created an awkward stop-and-start cycle . Contracts got longer and longer , until their duration outlasted the average tenure of CIOs .
Keeping the lights on
Perception of change
Companies with multiyear contracts — and the BPO providers who served them — could do little more than keep the lights on . Many BPO providers fell behind everywhere else . Keeping up with the basics was out of reach because even the volume demands of a basic corporate email system change significantly every year . Any efforts to meet new business needs stalled because organizations underestimated the rate at which software changes . Fixed price contracts with tight service legal agreements ( SLAs ) drove proper repairs and overhauls past the contract end date . The BPO industry became an expert in keeping systems crawling , touching nothing until it failed and designing the cheapest fixes possible .
The result was astoundingly diminished IT capabilities for some of the world ’ s largest organizations . “ For companies who are now coming off five-year IT outsourcing contracts , it ’ s likely they ’ ve been frozen in time , during one of the most disruptive times in technology ,” remarked cloud architecture strategist Adrian Cockcroft . Any years-old IT contract is inadequate today . Because many contracts were too amorphous , commercial teams trying to reduce expenditures were occupied in a game of corporate whack-a-mole , reducing costs in one area just to see them pop up elsewhere .
Change to keep the lights on
Change to keep pace with the basics
Change to meet the business need
20 % 30 % 30 % 20 %
80 %
20 %
8 DIGITECH Magazine Summer 2017
BREAKING FREE OF THE “IF IT AIN’T BROKE” IT MODEL By Ben Grinnell BREAKING FREE OF THE “IF IT AIN’T THE STATE OF THE INDUSTRY AND BROKE” IT MODEL HOW IT GOT THERE When Netflix moved to the cloud, it Innovation and constant improvement released a “chaos monkey” to wreak havoc are essential to the long-term success on its system. The chaos monkey is a script of enterprise IT. Over the last couple of meant to mimic the behavior of “a wild decades, however, many leaders decided monkey with a weapon in your data center that IT was not a core capability they (or cloud region) to randomly shoot down wanted to worry about. They outsourced instances and chew through cables.” 1 The it, optimizing for cost and reliability. That idea was to cause intentional failures in decision built up an entire industry and a controlled environment where Netflix’s led many cost-focused organizations to experts could comfortably monitor and delay modernizing their requirements. IT rep "G2g&7G'V7GW&RfV&VB6vVF6RF6G&7G0vF'W6W72&6W72WG6W&6r%GW&rV2W&F2FRWFfƗFV&fFW'2vW&R&VWBFRB&Vg&W6VFW"7&V6VBFR7FfGbFR62&w&v2&&W"FV7&V6VBFRFW&6W2FPWG&72B&V6&FVBG22BF2&w&FRWrFVFW'2fFV@g&7G'V7GW&RF2&7FfR&6FFRGW7G'F&BF&Vg&W6BFV'VখFVFgr&&V2v2BWrFF( 2FV"B7B&&FVB67BFW6rFPBvW'2fR&VVvF&6RbB&Gv&RfBFW77VV@FvFVFW&6W2vVVVR&W'G2FBBWV6W2fW&vVBG&vFBb76V&ǒƖR&&V2G&&FW"FFVFW'2g&WVVFǒWV7FVBFV"7WƖW'0FWFV7BWfW"6W"vVW76W2BWFfƗFF&FW2BW7FrWfV2@FW6R6Vb֖fƖ7FVBF67FW'2V&VBFR77VVBFBFR7WƖW"vVB&V6W6FFVFgvVƖ2B'VBB&Vg&W667G2fW"FRƖfRbFRWpW&W6ƖV76B6VB6FVR6W'fr6G&7B7W7FW'2vV&VfW&W267W'&VBखFR6'BFW&F2FVVV@FW&B6WF2W7B&R&RFF42&V6W6RBv2V76V6FVRgV7FrvVFV"vF&VvWBFR&v旦FfW"FPvB2'&VFFR77FV'f'7BF2bFR6G&7BWfW'Pv&r66VǒvFFW"FWfVW'2BvBWrFV6wBFW&旦V@6W'f6RvW'2FWfVW'2'VBFVW6W'f6W2BFR67BbBVB7FVGVFW'7FFrbrFV"Ɩ6F2Vf'GVFVǒF2FVFF( B66VBf Bg&7G'V7GW&RW&f&F&VvFR7&V6r6WGBGf6VV@7VGW&RbV7W&VVB6FVW2bBW6W'>( w&vrWV7FF2b&60&fVVBBV&rF22FRǒ6W'f6W2BFR&6rB&Vg&W667G0vF7W'ffRFRWfF&RF67FW'2FB66VBBF27&VFVBvv&B7FB7F'B76R6G&7G0vBvW"BvW"VFFV"GW&FWF7FVBFRfW&vRFVW&Rb42ࣀDtDT4vP7VW"#pUN( 2DRBDR$TBTdd%@6vRFVWFRƖvG26vRFVW6RvFFR&670VWrFPƖvG2#PW&6WF`6vP6vRFVW@FR'W6W72VV@3P3PP#RBTdd%@#R%U4U50@6W2vFVFV"6G&7G>( F@FR%&fFW'2v6W'fVBFV( F6V@FƗGFR&RFVWFRƖvG2琤%&fFW'2fV&VBWfW'vW&PV6RVWrWvFFR&672v0WBb&V6&V6W6RWfVFRfVPFVG2b&626'&FRV77FV6vR6vf6FǒWfW'V"琦Vff'G2FVWBWr'W6W72VVG27FV@&V6W6R&v旦F2VFW&W7FFV@FR&FRBv66gGv&R6vW2fV@&6R6G&7G2vFFvB6W'f6RVvw&VVVG242G&fR&W"&W'0BfW&V27BFR6G&7BVBFFRFR%GW7G'&V6RWW'@VWr77FV27&vƖrFV6pFrVFBfVBBFW6vrFP6VW7BfW276&RFR&W7VBv27FVFvǒF֖6VB@6&ƗFW2f"6RbFRv&N( 2&vW7@&v旦F2( f"6W2v&Pr6֖rfbffRזV"BWG6W&6p6G&7G2N( 2ƖVǒFW( fR&VVg&VখFRGW&rRbFR7BF7'WFfPFW2FV6w( &V&VB6V@&6FV7GW&R7G&FVv7BG&667&gBV'2BB6G&7B2FWVFPFF&V6W6R6G&7G2vW&RF'W26W&6FV2G'rF&VGV6RWVFGW&W2vW&R67WVBvRb6'&FRv6R&VGV6p67G2R&VW7BF6VRFVWV6WvW&R