Towards an Interoperable Scientific Cloud for Europe

To ensure world-class research, energy effi­cien­cies and com­pet­i­tive edge in the global mar­ket­place, Europe needs to evolve cur­rent Dis­trib­uted Com­put­ing Infra­struc­tures (DCIs) that encom­pass new, industrial-quality tech­nolo­gies such as vir­tu­al­iza­tion, ser­vice ori­en­ta­tion and con­ver­gence with the dig­i­tal world. While grid infra­struc­tures have cap­tured the require­ments of sev­eral spe­cific com­mu­ni­ties, smaller and ad-hoc groups with sig­nif­i­cant appli­ca­tions have strug­gled to get their require­ments sat­is­fied with grid tech­nol­ogy because of the inher­ent com­plex­ity and long deploy­ment times (with out­comes not always meet­ing with success).

More­over, indus­try adop­tion of grid has not taken off as widely as once expected. By con­trast, a busi­ness case for cloud com­put­ing is increas­ingly gain­ing con­sen­sus in both the pub­lic and pri­vate sec­tors and as sev­eral stan­dard­i­s­a­tion devel­op­ment organ­i­sa­tions focus efforts on inter­op­er­a­ble solu­tions for clouds through strate­gic alliances in which Europe is play­ing a pro-active role. Fur­ther­more, a recent Expert Group Report on the Future of Cloud Com­put­ing pro­duced with the sup­port of the Euro­pean Com­mis­sion DG INFSO rec­om­mends that the Euro­pean open source move­ment should work strongly with indus­try to sup­port com­mer­cial cloud based ser­vice provisioning.

A cloud-based e-Infrastructure for eScience, cur­rently miss­ing from Europe’s ser­vice port­fo­lio, would ensure a leap for­ward in the Euro­pean Research Area by inte­grat­ing flex­i­ble and easy-to-use util­ity ser­vices, com­ple­ment­ing cur­rent com­put­ing ser­vices like grids and super­com­put­ers at the hands of researchers and sci­en­tists. Value-add needs to come from new busi­ness mod­els in a shift away from costly and com­plex “run-by-scientists-for-scientists” approaches on the one hand and the use of pay on demand on the other. Sus­tain­able growth needs to be addressed by a deeper under­stand­ing of pol­icy and legal issues, ensur­ing cost-effective invest­ment at EU level and inter­op­er­abil­ity while also fos­ter­ing new public-private part­ner­ships in the longer term. A new cul­ture of cloud research, “sci­en­tific cloud”, and a spirit of entre­pre­neur­ship can­not be achieved with­out the involve­ment in R&D ini­tia­tives of pio­neer­ing enter­prises with a com­mit­ment to indus­try qual­ity stan­dards and inter­op­er­abil­ity work­ing along­side research organisations.

Recent devel­op­ments led by experts in indus­try and research would help to gain effi­cien­cies and make sav­ings by opti­mis­ing resource util­i­sa­tion, reli­a­bil­ity, energy effi­ciency and main­te­nance costs, all key objec­tives high­lighted by EU pol­icy bod­ies. This new approach focuses on the pro­vi­sion­ing, oper­a­tion and user-testing of an indus­trial qual­ity, vir­tu­alised e-Infrastructure in the form of a cloud com­put­ing ser­vice plat­form, open for usage by the research and sci­en­tific com­mu­nity and tested by major cat­e­gories of sci­en­tific and indus­trial com­mu­ni­ties across dis­ci­plines and sec­tors impor­tant to Europe. The aims of these new devel­op­ments are to broaden inter-disciplinary sci­en­tific col­lab­o­ra­tion in Europe, ensure co-ordinated, strength­ened and focused soft­ware deploy­ments, improve the usabil­ity of DCI plat­forms tar­get­ing the largest pos­si­ble base across a range of fields in sci­ence and engi­neer­ing, and advance exploita­tion in the rapidly chang­ing hard­ware envi­ron­ments through appro­pri­ate soft­ware developments.

This novel com­po­nent in the e-Infrastructure ecosys­tem would help expand exist­ing Dis­trib­uted Com­put­ing Infra­struc­tures (DCIs) serv­ing eScience by ensur­ing easy access to vir­tu­ally “infi­nite” resources and high mobil­ity while hid­ing the com­plex­ity of set-up, main­te­nance and com­mu­ni­ca­tion from users and reduc­ing the length and costs of appli­ca­tion port­ing through automa­tion, as well as over­com­ing the need for in-depth knowl­edge of ICT tech­nolo­gies. Economies of scale will be achieved by opti­mis­ing resources, reduc­ing oper­a­tional costs, espe­cially energy costs, where sav­ings are cru­cial for sustainability.

An ideal approach could be based on both open source and com­mer­cial solu­tions, com­bin­ing the best of both worlds. Users would be enabled through access to a com­mer­cial multi-layer solu­tion includ­ing com­pute and stor­age power, a devel­op­ment envi­ron­ment and imme­di­ate ser­vices, while advances in open source would also be ensured through com­mu­nity con­tri­bu­tions to extend the capa­bil­i­ties of cur­rent DCIs and sup­port efforts towards inter­op­er­abil­ity and portability.

Open source ini­tia­tives would be lever­aged to pave the ground for inter­op­er­abil­ity. A good case in point is the Zend Frame­work project, which has invited the open source com­mu­nity and soft­ware ven­dors to par­tic­i­pate in the for­ma­tion of a Sim­ple Cloud API. IBM, Microsoft, Rack­space, Nir­vanix and GoGrid have already joined the project as con­trib­u­tors. In com­ing months, they will work together to define APIs for these cloud appli­ca­tion ser­vices, enabling a gen­er­a­tion of cloud native appli­ca­tions writ­ten in PHP . The Sim­ple Cloud API is an open source project that makes it eas­ier for devel­op­ers to use cloud appli­ca­tion ser­vices by abstract­ing insignif­i­cant API dif­fer­ences. One of the design goals of the project is to encour­age inno­va­tion. To this end, the Sim­ple Cloud API can be used for com­mon oper­a­tions while users can eas­ily drop down to ven­dor libraries to access value-add fea­tures. One exam­ple of this is Microsoft Azure, which now also sup­ports the full Java stack includ­ing open source tools such as the Apache web server, work­ing towards interoperability.

But it doesn’t stop here. A cost and energy effi­cient on-demand envi­ron­ment has much poten­tial to sup­port incu­ba­tors, indus­trial clus­ters and sci­en­tific parks, which are cen­tral to Europe’s eco­nomic strength, par­tic­u­larly in terms of high value-added cat­e­gories like ICT, Biotech­nol­ogy and Phar­ma­ceu­ti­cals and R&D across diverse sec­tors. What’s more, such a solu­tion would enable SME and small research labs by bring­ing the value-add needed to com­pete with the larger organ­i­sa­tions that cur­rently dom­i­nate the phar­ma­ceu­ti­cal landscape.

Sig­nif­i­cantly, such an approach meets with all four addi­tional rec­om­men­da­tions of the EC’s Expert Group Report for the future of cloud com­put­ing, that is, the need for large-scale research and exper­i­men­ta­tion test beds; devel­op­ing joint pro­grammes encour­ag­ing expert col­lab­o­ra­tion groups with indus­trial and pub­lic stake­hold­ers; sup­port­ing the devel­op­ment of cloud inter­op­er­a­tion stan­dards and open source ref­er­ence imple­men­ta­tion and Euro­pean lead­er­ship posi­tion in soft­ware through com­mer­cially rel­e­vant open source approaches. The time has come for Europe to tap into the exper­tise that will help make this hap­pen, open­ing up strate­gic oppor­tu­ni­ties for a new sci­en­tific cloud that brings inter­op­er­abil­ity and inno­va­tion into sharp relief.

—–

Source: Trust-IT Ser­vices Ltd.
via HPCwire: Towards an Inter­op­er­a­ble Sci­en­tific Cloud for Europe.

Tags: , ,

Be the first to comment on "Towards an Interoperable Scientific Cloud for Europe"

Leave a Comment

*

Get Adobe Flash player