For Hy­per-Con­verged In­te­grated Sys­tems

Dataquest - - INDUSTRY -

Gart­ner, Inc. has iden­ti­fied seven of the most com­mon flawed as­sump­tions in the hy­per­con­verged in­te­grated sys­tems (HCIS) mar­ket “HCIS, which en­com­passes soft­ware­cen­tric ar­chi­tec­tures that in­te­grate com­pute, stor­age and net­work­ing on com­mod­ity hard­ware, prom­ises a cost-ef­fec­tive in­fra­struc­ture so­lu­tion that is sim­ple to de­ploy, man­age and scale,” said Ge­orge Weiss, Vice Pres­i­dent and dis­tin­guished an­a­lyst at Gart­ner. “How­ever, new and emerg­ing tech­nolo­gies are of­ten sur­rounded by hype as ven­dors try to ac­cel­er­ate sales. In­fra­struc­ture and Op­er­a­tions (I&O) lead­ers and de­ci­sion mak­ers should ex­am­ine the fol­low­ing points care­fully to avoid later dis­ap­point­ments or traps.”

Myth 1: All im­ple­men­ta­tions com­prise stan­dard and open ar­chi­tec­tures

In the soft­ware-de­fined world of HCIS, the lev­els of stan­dard­iza­tion and open­ness de­pend in­creas­ingly on the code base. It’s im­por­tant to be clear on who con­trols the code and who is re­spon­si­ble for its de­vel­op­ment, main­te­nance and per­for­mance. There are no soft­ware-de­fined stan­dards, so one ven­dor’s man­age­ment con­trols may not man­age an­other ven­dor’s de­vices or soft­ware-de­fined net­work.

Myth 2: All im­ple­men­ta­tions are des­tined to fail mis­sion-crit­i­cal scal­a­bil­ity and re­siliency tests

HCIS im­ple­men­ta­tions will vary widely in ro­bust­ness, scal­a­bil­ity and se­cu­rity. The vi­tal con­text that needs to be con­sid­ered is the in­tended use case. HCIS is de­signed and best-suited to high-avail­abil­ity and vir­tu­al­ized work­loads. Yet even here, there is wide vari­ance; some HCIS clus­ters scale only to eight nodes, whereas oth­ers claim to scale to hun­dreds or even thou­sands. “Cau­tion is ad­vised as this kind of scal­a­bil­ity of­ten does not fit the Gart­ner def­i­ni­tion of a seam­lessly man­aged HCIS ap­pli­ance,” Mr. Weiss said.

Myth 3: HCIS costs rep­re­sent the least-ex­pen­sive de­ploy­ment model

HCIS in­fra­struc­ture can be scaled up eas­ily in small in­cre­men­tal ad­just­ments by adding ad­di­tional nodes as needed. How­ever, over an ex­tended pe­riod of time where the use-case de­mand rises and reg­u­larly re­quires ad­di­tional nodes, the in­vest­ment in HCIS could eas­ily ex­ceed an up­front in­vest­ment. 74 www.dqin­dia.com

Myth 4: The most im­por­tant use case is Vir­tual Desk­top In­fra­struc­ture (VDI)

VDI has be­come the “celebrity” use case for HCIS. How­ever, many gen­eral-pur­pose work­loads are now a match for HCIS due to im­proved per­for­mance, scal­ing, data pro­tec­tion and ease of de­ploy­ment, as well as an ex­pand­ing hy­brid cloud ecosys­tem. I&O lead­ers should ex­pect fur­ther ex­pan­sion to oc­cur in the next three years to han­dle greater lev­els of agility, DevOps, con­tain­ers, bi­modal ap­pli­ca­tions and con­sumer-based ser­vices.

Myth 5: HCIS spells the demise of tra­di­tional stor­age ar­rays

HCIS has huge po­ten­tial to re­place small to mid-size, gen­eral-pur­pose disk ar­rays in highly vir­tu­al­ized en­vi­ron­ments. In the case of large mis­sion-crit­i­cal ap­pli­ca­tions that re­quire pre­dictable be­hav­ior and proven re­li­a­bil­ity, HCIS may be less ef­fec­tive. Once all ca­pac­ity uti­liza­tion and cost fac­tors are con­sid­ered, mod­ern hy­brid or solid­state ar­ray de­ploy­ments are likely to be more eco­nom­i­cal over the long haul.

Myth 6: HCIS elim­i­nates data cen­ter in­ter­op­er­abil­ity and si­los

HCIS lacks tight in­te­gra­tion with ex­ist­ing tra­di­tional in­fras­truc­tures, which forces I&O lead­ers to po­si­tion them in silo de­ploy­ments. The silo ap­proach ac­com­mo­dates the ex­ist­ing ad­min­is­tra­tion and tech­ni­cal sup­port of legacy op­er­a­tions. Nev­er­the­less, HCIS also de­mands new mod­els of team col­lab­o­ra­tion and spe­cial­ity in­te­gra­tions that are dif­fer­ent to ex­ist­ing legacy so­lu­tions. HCIS de­ploy­ment mod­els res­onate most with IT lead­ers who want to switch from hard­ware stack man­age­ment mod­els to sim­ple-to-de­ploy vir­tu­al­ized plat­form de­liv­ery.

Myth 7: Tra­di­tional ven­dor se­lec­tion pref­er­ence will re­main the same

Gart­ner’s fo­cus group par­tic­i­pants have shown that loy­alty to tra­di­tional ven­dors would be tested by sev­eral cri­te­ria:

Is the ven­dor in­creas­ingly flu­ent in the new wave of HCIS?

Is it se­ri­ously will­ing to dis­rupt so­lu­tions? Does it have the vi­sion to drive in­no­va­tion? Can it keep ahead of emerg­ing, ag­ile com­peti­tors and in­crease its sav­ings? its con­ven­tional

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.